basement geochemistry and geochronology of central malaita

JOURNAL OF PETROLOGY VOLUME 43 NUMBER 3 PAGES 000–000 2002

Basement Geochemistry and Geochronologyof Central Malaita, Solomon Islands, withImplications for the Origin and Evolution ofthe Ontong Java Plateau

M. L. G. TEJADA1∗, J. J. MAHONEY2, C. R. NEAL3, R. A. DUNCAN4

AND M. G. PETTERSON5

1NATIONAL INSTITUTE OF GEOLOGICAL SCIENCES, UNIVERSITY OF THE PHILIPPINES, DILIMAN, QUEZON CITY,

1101 PHILIPPINES2SCHOOL OF OCEAN AND EARTH SCIENCE AND TECHNOLOGY, UNIVERSITY OF HAWAII, HONOLULU, HI 96822, USA3DEPARTMENT OF CIVIL ENGINEERING AND GEOLOGICAL SCIENCES, UNIVERSITY OF NOTRE DAME, NOTRE DAME,

IN 46556, USA4COLLEGE OF OCEANIC AND ATMOSPHERIC SCIENCES, OREGON STATE UNIVERSITY, CORVALLIS, OR 97331, USA5BRITISH GEOLOGICAL SURVEY, ONSHORE MINERALS AND ENERGY RESOURCES, KEYWORTH NG12 5GG, UK

RECEIVED OCTOBER 10, 2000; REVISED TYPESCRIPT ACCEPTED SEPTEMBER 7, 2001

KEY WORDS: Ontong Java Plateau; geochemistry; mantle sources; petro-Sections of Ontong Java Plateau basalt basement in central Malaitagenesis; eruptive stratigraphy(Solomon Islands) are 0·5–3·5 km thick and resemble a much-

expanded version of that recovered at Ocean Drilling Program Site807. 40Ar–39Ar ages (121–125 Ma) are identical to those for Site807, southern Malaita, Ramos Island, parts of the island of SantaIsabel, and Deep Sea Drilling Project Site 289; the >90 Ma INTRODUCTIONeruptive episode seen in Santa Isabel, San Cristobal, and at drill The Solomon islands of Malaita, Ramos, Ulawa, thesites 803 and 288 is not represented. The central Malaitan basalts northeastern part of Santa Isabel, and probably part ofprovide further evidence of two distinct ocean-island-like mantle San Cristobal (a.k.a. Makira) form the obducted andsources, and the combined data preclude a significant contribution uplifted southwestern margin of the world’s largest vol-from normal ocean-ridge-type mantle. As at Site 807, two geo- canic oceanic plateau, the 5 × 107 km3 Ontong Javachemically distinct stratigraphic groups are present, the Singgalo Plateau (OJP; Fig. 1) (e.g. Kroenke, 1972; Hughes &Formation (>750 m thick in central Malaita) and the Kwaimbaita Turner, 1977; Coleman et al., 1978; Coleman & Kroenke,Formation (>2700 m thick). Both a peridotite plume-head and 1981; Petterson et al., 1997; Phinney et al., 1999). Re-eclogite-bearing plume-head may account for the geochemical char- connaissance studies show that the Cretaceous igneousacteristics, but the observed stratigraphic succession requires special basement of these islands is almost indistinguishable inconditions for the latter model. A number of first-order features of age from, and closely similar in composition to, thatthe Ontong Java Plateau do not obviously fit the predictions of any drilled at Ocean Drilling Program (ODP) Sites 807 andplume-head model: for example, at least two important, geochemically 803 and Deep Sea Drilling Project (DSDP) Site 289 onsimilar eruptive episodes >30 my apart, the lack of an obvious the northern and north–central plateau (Mahoney et al.,plume-tail trace, and lack of evidence for emergence or significant 1993; Parkinson et al., 1996; Tejada et al., 1996; Neal et

al., 1997). Recently, we sampled accessible portions ofuplift.

∗Corresponding author. Telephone: 632-435-4837. Fax: 632-929-6047.E-mail: [email protected] Oxford University Press 2002

JOURNAL OF PETROLOGY VOLUME 43 NUMBER 3 MARCH 2002

0·5- and 3·5-km-thick sections of OJP crust in central Our principal study area was the northern part of theKwaio Anticline, where the thickest exposures of theMalaita. As the deepest exposures of the plateau’s crust

yet sampled systematically, these sections present a unique Malaita Volcanic Group occur (Fig. 2a). The anticlineis accessible via rivers that cut the basement and sedi-source of information on age, eruptive stratigraphy, petro-

genesis, and mantle sources. In this paper, we characterize mentary sections in a SW–NE direction roughly per-pendicular to the axis of the anticline, facilitating thicknessthe geochemical stratigraphy of the southern OJP crust

recorded in these sections, and build upon previous work determination of all the lithostratigraphic units. Themajority of the samples we discuss here were collectedto examine what the isotopic and chemical variations

reveal about the plateau’s origin and the implications for from the Singgalo and Kwaimbaita rivers, where a lavasection of estimated 3·5 km thickness (Petterson et al.,the mantle source regions that fed plateau magmatism.1997) forms the core of the anticline (Fig. 2b). The othersample area we discuss is the Kwara’ae Anticline (Fig.2a), accessible via the Kwaiafa’a River and its tributary,STUDY AREA AND GENERAL FIELDthe Bisula, where we sampled a basement section of

CHARACTERISTICS OF OJP CRUST >500 m thickness.Samples were taken mainly from interior portions ofIN MALAITA

pillows and massive flows, to minimize alteration. TheCretaceous igneous OJP crust forms the cores of severalmodal mineralogy of the microgabbros and the ground-large anticlines in Malaita (Fig. 2a) and is exposed inmass in the aphyric and plagioclase-phyric basalts consistsstream gullies and waterfalls on this heavily vegetatedof almost equal proportions of unaltered plagioclase (upisland. The basement rocks, formerly known as the Olderto 58%) and clinopyroxene (up to 45%), with minorSeries or Older Basalts in the south (Hughes & Turner,altered olivine (5–7%) and oxides (Petterson, 1995).1976, 1977), the Fiu Lavas in the north (Rickwood, 1957),Plagioclase and clinopyroxene are usually in an ophiticand the Older Malaita Volcanics in the northernmost partto subophitic relationship, indicating contemporaneousof the island (Barron, 1993), recently have been termedcrystallization, although in some sections plagioclase crys-the Malaita Volcanic Group (Petterson et al., 1997). Thetals are found included in clinopyroxene grains, possiblyMalaita Volcanic Group is distinguished in the fieldindicating that the plagioclase crystallized earlier. Modalfrom an overlying, much less extensive volcanic series,abundance of mesostasis varies from 10 to 30% butpreviously called the Younger Series or Younger Basaltsusually is <20%. In the samples chosen for geochemical(Hughes & Turner, 1976, 1977) but now termed thework, limited low-grade alteration is seen in rare amyg-Maramasike Formation (Fm.; Petterson et al., 1997), bydules containing zeolite, chlorite, and calcite, and isthe vesicularity of the younger volcanic rocks and themanifested in mesostasis and olivine, which are partiallyintercalation of these rocks with Eocene pelagic lime-to wholly replaced by brown clay and iddingsite (Babbs,stones.1997). In some of the more altered samples (e.g. ML476)Samples for this study were collected in conjunctionplagioclase is partially replaced by a mixture of clay andwith geological mapping surveys by the Solomon Islandscalcite, whereas in others (e.g. SGB24), secondary phasesMinistry of Energy, Mineral and Water Resources inreplace olivine completely and fill the interstices between1993 (see Petterson, 1995). The great bulk of the Malaitafeldspar laths. In the Kwaio Anticline, a minor butVolcanic Group exposed in the central part of the islandstriking facies of the Malaita Volcanic Group, consistingconsists of massive and pillowed lava flows very similarof basalts with ‘spherulitic’ or ‘orbicular’ texture, wasto those in southern Malaita, and to the Sigana Basaltsfound as boulders in the river beds (Petterson, 1995).(Hawkins & Barron, 1991) in northeastern Santa Isabel.This facies consists of 1–6 cm fragments of ophitic gabbroThe flows are aphanitic to phaneritic, aphyric to plagio-and plagioclase or clinopyroxene megacrysts in basalticclase + clinopyroxene (± olivine)-phyric basalts, andto diabasic host rock.coarser-grained diabases and microgabbros [see Petterson

(1995) and Babbs (1997) for detailed descriptions]. Im-portant findings of the surveys are that all of the lavas wereemplaced in a submarine environment, that interbeds of

ANALYTICAL METHODSchert or mudstone are very rare, and that no limestoneWe collected 152 igneous rock samples, 64 of which weinterbeds were found. Very few dikes were encountered,chose for geochemical study: 30 from the Singgalo River,although gabbroic intrusions were found to the north,20 from the Kwaimbaita River, and 14 from the Kwai-within the Fateleka Anticline (Fig. 2a). Except in disturbedafa’a River drainages. Samples were trimmed with asections, the basalts in central Malaita are everywherewater-cooled rock saw to remove alteration rinds andoverlain conformably by the early to middle Cretaceoussplit into three parts; the freshest interior portions wereKwara’ae Mudstone Formation, consisting of siliceous,

parallel-laminated radiolarian pelagic mudstones. reserved for 40Ar–39Ar and isotopic analysis. Splits for

2

TEJADA et al. ORIGIN AND EVOLUTION OF ONTONG JAVA PLATEAU

Fig. 1. Etopo 5 predicted bathymetry (in meters) of the western Pacific Basin (after Smith & Sandwell, 1996), centered on the Ontong JavaPlateau. The Solomon Islands are at the plateau’s southwestern margin. Also shown are the locations of ODP Sites 807, 803, and 289, wherebasement rocks were recovered, and 288, which recovered > 90 Ma and >118 Ma ash layers but did not reach basement.

bulk-rock major and trace element analysis were chipped McGinnis et al. (1997); see Jain & Neal (1996) for details.Of the elements analyzed by both XRF and ICP-MS,into pieces 0·3–0·5 cm across, cleaned ultrasonically in

deionized water, and dried; chips showing macroscopic the agreement is generally (Ζ5% for Zr and (Ζ15% forY; agreement is within 15% for Co and Nb and 20%alteration significantly greater than the average were

removed, and the remainder powdered in an alumina for Cr (except for one, four, and six samples, respectively).The ICP-MS values for Rb are systematically highermill. Major elements and a suite of trace elements (Table

1; note that samples are listed in stratigraphic order from than the XRF values by as much as 0·9 ppm; however,the analytical uncertainty of the XRF measurements istop to bottom of each section) were analyzed by X-ray

fluorescence (XRF) spectrometry at the University of relatively large as Rb contents are generally near theXRF detection limit.Hawaii, using the methods of Norrish & Chappell (1977).

Rare earth elements and several other trace elements Splits of eight whole-rock samples were prepared atOregon State University for 40Ar–39Ar incremental-heat-(Table 2) were analyzed on 0·2 g splits of the powders

by inductively coupled plasma–mass spectrometry (ICP- ing age determinations. Small cores (>0·7 g) cut fromthe freshest portions of samples were loaded in evacuatedMS) at the University of Notre Dame using the modi-

fication of Jenner et al.’s (1990) method proposed by quartz vials and irradiated for 6 h at 1 MW power at

3


Fig. 2. (a) Simplified geologic map of the northern half of Malaita showing the anticlines where exposures of the Malaita Volcanic Group arefound (after Petterson, 1995). The rivers traversed are shown with heavier-outlined portions indicating sampling coverage. The line A–A’represents the projection line for the cross-section in (b). (b) Basement structure along the Singgalo and Kwaimbaita rivers (after Petterson et al.,1997); the boundary between the two chemical groups of lavas we term the Singgalo Fm. and Kwaimbaita Fm. (see text) is determined fromisotopic and chemical data. The arrows at the bottom indicate flow directions of the Singgalo and Kwaimbaita rivers, and the filled barsrepresent our sampling coverage. The open bar along the upper Singgalo River shows the extent of sampling conducted by T. Babbs andA. Saunders (Babbs, 1997).

the Oregon State University TRIGA reactor. Biotite mean of concordant, adjacent step ages), isochron ages(determined from the linear regression of step com-standard FCT-3 (27·55 ± 0·12 Ma, equivalent to

513·9 Ma for hornblende Mmhb-1; Lanphere et al., 1990) positions), and integrated or total fusion ages (calculatedby adding all step compositions together).was used to monitor the neutron fluence. The isotopic

composition of Ar was determined for each of 4–6 Sample preparation and analysis for Sr, Nd, and Pbisotopes were performed at the University of Hawaii [fortemperature steps using an AEI MS-10S mass spec-

trometer [see Duncan et al. (1997) for experimental methods, see Peng & Mahoney (1995) and Tejada et al.(1996)]. Isotope-dilution abundance data for Sr, Rb, Nd,details]. Table 3 includes plateau ages (the weighted

4

TE

JAD

Aet

al.O

RIG

INA

ND

EV

OL

UT

ION

OF

ON

TO

NG

JAV

APL

AT

EA

U

Table 1: X-ray fluorescence major and trace element data for central Malaita basement rocks

Sample SiO2 TiO2 Al2O3 Fe2O3∗ MnO MgO CaO Na2O K2O P2O5 SUM LOI Rb Nb Sr Zr V Y Sc Zn Cu Co Mn Ni Cr

Kwaio Anticline

Singgalo River

SG1 50·38 1·65 13·54 14·01 0·19 7·30 11·64 2·32 0·15 0·14 101·32 0·30 1·8 4·4 136 81 392 29 41 93 124 61

SG2 49·65 1·59 14·15 13·68 0·18 7·69 11·25 2·30 0·16 0·14 100·79 0·71 1·9 4·3 141 78 372 25 38 87 99 85

SG3 49·75 1·52 14·31 13·82 0·18 7·59 11·14 2·30 0·17 0·15 100·93 0·67 1·8 4·9 144 88 329 28 34 90 97 74

SG4 49·98 1·62 14·22 13·85 0·19 6·99 11·34 2·38 0·11 0·15 100·83 0·51 <1·5 5·4 143 93 395 28 41 99 96 62

SG5 49·73 1·57 14·70 12·54 0·18 7·44 11·29 2·43 0·19 0·15 100·22 0·89 1·9 5·2 145 91 395 28 44 96 94 71

SG6 49·50 1·57 14·36 12·98 0·19 7·01 12·52 2·42 0·16 0·14 100·85 1·56 <1·5 4·5 146 80 384 28 39 89 103 86

SG7 50·18 1·67 14·19 13·83 0·19 6·86 11·28 2·48 0·14 0·16 100·98 0·42 <1·5 5·5 145 98 376 31 42 102 92 70

SGB1 49·59 1·68 13·92 13·92 0·18 7·04 10·86 2·44 0·18 0·15 99·96 0·74 1·6 5·5 139 92 383 29 43 96 96 67

SGB2 49·96 1·60 14·47 13·06 0·19 6·88 11·23 2·52 0·20 0·16 100·27 1·24 1·9 5·8 145 94 397 30 49 104 93 73

SGB6 50·07 1·68 14·33 13·29 0·19 7·02 11·11 2·25 0·15 0·14 100·23 0·28 <1·5 5·6 146 97 390 30 42 118 109 52 1440 79 123

SGB7 49·83 1·64 14·11 13·93 0·19 7·04 11·41 2·00 0·16 0·13 100·44 0·71 2·1 4·7 146 90 342 30 36 97 107 47 1290 73 107

SGB8 50·39 1·77 14·33 13·63 0·18 6·87 10·93 2·26 0·16 0·14 100·66 0·75 <1·5 5·6 148 99 367 31 36 110 116 50 1230 71 90

SGB3 49·45 1·62 13·98 14·24 0·20 6·34 11·52 2·35 0·11 0·15 99·96 0·52 <1·5 5·5 146 96 353 30 39 102 94 71

SGB4 49·62 1·62 14·05 14·33 0·20 6·72 11·46 2·40 0·13 0·16 100·69 0·71 <1·5 5·8 145 95 351 30 38 100 92 72

SGB5 49·89 1·40 14·48 13·71 0·18 6·95 11·59 1·82 0·16 0·13 100·31 1·02 1·7 5·3 151 93 327 29 31 103 113 48 1160 76 80

SGB9 49·39 1·49 14·51 13·14 0·20 7·44 11·88 2·09 0·14 0·12 100·40 0·24 2·0 4·2 152 78 308 26 33 86 115 46 1380 87 154

SGB12 48·82 1·53 15·13 12·91 0·21 7·14 11·31 2·15 0·14 0·15 99·49 1·08 <1·5 5·6 150 99 427 31 50 111 109 54 1640 77 153

SGB10 49·92 1·35 14·01 14·37 0·20 6·81 10·95 2·21 0·17 0·14 100·13 0·46 2·6 6·0 141 92 360 30 38 106 102 47 1380 64 100

SGB11 49·11 1·39 14·24 14·26 0·21 7·15 11·64 2·19 0·08 0·15 100·42 0·17 <1·5 5·2 147 96 367 31 38 105 108 50 1520 78 118

SGB13 50·00 1·48 14·17 13·73 0·20 6·86 11·14 2·30 0·16 0·15 100·19 0·45 <1·5 5·8 148 100 349 32 38 104 105 49 1430 72 106

SGB14 49·70 1·50 14·09 14·26 0·20 6·94 11·23 2·21 0·18 0·14 100·45 0·68 2·2 5·5 148 96 319 31 32 129 110 49 1370 74 91

SGB16 49·05 1·37 14·19 14·02 0·19 7·34 11·45 2·16 0·11 0·14 100·02 1·04 <1·5 5·0 147 94 320 30 31 95 115 50 1240 85 95

SGB15 49·39 1·13 14·46 13·04 0·19 7·35 12·25 2·05 0·12 0·11 100·09 0·78 <1·5 4·3 143 76 307 25 32 88 128 48 1360 86 110

SGB17 49·48 1·18 14·11 13·50 0·21 7·14 11·99 2·02 0·14 0·12 99·89 0·50 1·7 4·5 141 83 326 27 32 101 129 49 1430 76 84

SGB18 49·01 0·88 14·72 12·72 0·18 7·42 13·03 1·81 0·10 0·10 99·97 0·51 <1·5 3·3 121 66 300 23 41 87 142 53 1360 99 198

SGB22 48·93 0·98 14·65 12·86 0·19 7·41 12·95 1·84 0·09 0·10 100·00 0·35 <1·5 3·1 124 67 318 23 42 91 145 53 1490 99 184

SGB23 49·00 1·05 14·37 13·10 0·19 7·54 12·77 1·83 0·10 0·10 100·05 0·63 <1·5 3·8 123 69 314 24 37 90 150 54 1320 96 159

SGB24 48·99 0·93 14·79 12·19 0·18 8·19 11·77 2·84 0·31 0·08 100·27 2·70 2·6 2·4 219 48 250 19 35 64 139 50 1140 101 189

SGB25 48·96 0·93 14·78 12·56 0·21 7·86 13·16 1·51 0·05 0·10 100·12 0·49 <1·5 3·6 122 65 320 22 45 92 141 52 1660 105 213

5

JOU

RN

AL

OF

PET

RO

LO

GY

VO

LU

ME

43N

UM

BE

R3

MA

RC

H2002

Table 1: continued


Kwaimbaita River

ML407 49·67 1·35 13·64 13·47 0·19 7·45 11·38 3·08 0·26 0·13 100·62 2·14 2·6 3·8 291 71 295 25 40 57 229 62

ML476 47·88 2·07 13·21 18·49 0·25 5·83 9·61 3·08 0·09 0·17 100·68 3·62 <1·5 5·5 129 101 407 33 31 111 174 35

ML475 49·68 0·73 14·00 9·84 0·16 9·99 14·52 1·48 0·08 0·06 100·54 1·91 <1·5 1·6 128 28 217 13 40 47 128 133

ML474 48·60 1·33 14·75 13·03 0·22 7·10 12·63 2·13 0·05 0·13 99·97 1·12 <1·5 4·2 155 73 324 25 45 89 137 104

ML436 49·06 1·27 14·55 12·41 0·22 7·82 12·67 2·18 0·14 0·12 100·44 0·94 <1·5 3·8 139 71 325 24 47 86 130 110

ML434 49·28 1·25 14·40 12·41 0·22 8·00 12·57 2·14 0·21 0·13 100·61 1·14 <1·5 3·9 138 69 316 23 45 81 124 103

ML432 49·30 1·24 14·69 12·35 0·24 8·05 12·66 1·91 0·12 0·12 100·68 2·51 1·6 3·2 180 64 314 22 37 68 143 99

ML431 49·05 1·30 14·60 12·63 0·20 7·80 12·87 2·00 0·06 0·12 100·63 0·71 <1·5 3·5 121 66 305 23 42 84 137 97

ML423 48·96 1·22 14·67 12·86 0·21 7·40 13·22 2·05 0·05 0·12 100·76 0·50 <1·5 3·4 122 64 296 22 41 77 133 97

ML422 49·24 1·22 14·75 11·93 0·26 7·95 12·87 2·14 0·06 0·12 100·54 0·45 <1·5 3·7 154 69 328 24 45 90 143 95

ML451 48·96 1·24 14·39 13·06 0·19 8·93 11·96 1·87 0·10 0·12 100·82 1·55 <1·5 3·2 129 62 316 22 29 78 150 125

ML450 49·50 1·60 14·28 13·37 0·19 7·42 11·66 2·28 0·14 0·14 100·58 0·94 2·0 4·5 133 81 371 27 37 82 109 74

ML449 50·09 1·69 14·80 13·04 0·19 7·25 10·42 2·45 0·19 0·16 100·28 0·95 1·7 5·7 146 95 412 30 48 101 100 76

ML421 49·17 1·76 14·18 13·94 0·19 7·66 11·06 2·34 0·07 0·16 100·53 0·84 <1·5 5·7 146 96 405 29 42 105 98 67

ML437 49·33 1·59 14·39 13·62 0·22 7·78 11·25 2·27 0·17 0·14 100·76 0·93 2·0 4·8 135 83 367 27 38 90 112 75

ML438 49·76 1·59 13·87 14·23 0·19 7·20 10·75 2·33 0·18 0·15 100·25 0·80 2·1 5·0 142 92 374 29 36 118 104 54

ML439 49·00 1·70 14·49 13·76 0·25 7·21 11·50 2·46 0·09 0·16 100·62 0·38 <1·5 5·7 149 100 409 31 49 109 95 62

ML440 49·60 1·36 14·09 13·56 0·18 7·65 11·68 2·26 0·15 0·14 100·67 0·72 2·0 4·6 131 80 321 26 33 81 111 80

ML448 50·31 1·61 14·41 12·92 0·22 7·89 10·51 2·44 0·18 0·16 100·65 0·99 <1·5 5·4 148 94 410 33 43 119 97 74

ML468 48·86 1·71 14·32 14·58 0·20 7·56 10·85 2·26 0·08 0·16 100·58 1·07 <1·5 5·8 148 95 401 29 45 109 96 60

Kwara’ae Anticline

Kwaiafa’a River

KF-1 50·13 1·28 14·35 12·23 0·19 8·09 10·90 2·94 0·23 0·08 100·41 1·94 3·0 3·5 127 59 293 20 31 70 117 55 1213 96 155

KF-4 50·47 1·72 13·77 14·22 0·21 6·60 11·03 2·03 0·13 0·14 100·31 0·50 2·2 5·7 152 101 441 31 34 102 115 41 1448 40 29

KF-9 49·79 1·56 14·06 14·13 0·21 6·39 11·50 1·98 0·15 0·14 99·91 0·49 2·1 5·5 144 97 334 30 37 102 104 49 1470 71 116

KF-13 48·63 1·59 14·55 14·35 0·21 6·80 11·88 1·65 0·04 0·14 99·83 0·42 <1·5 5·8 147 102 352 32 36 107 111 53 1578 80 115

KF-14 49·73 1·56 14·20 13·81 0·19 6·86 11·44 2·04 0·05 0·13 100·01 0·30 <1·5 5·5 146 96 343 31 35 106 105 48 1346 69 111

KF-19 49·48 1·62 14·11 14·02 0·20 7·02 11·18 1·94 0·08 0·12 99·77 0·41 <1·5 4·6 145 88 379 27 38 103 115 49 1474 69 116

KF-24 48·81 1·63 14·40 13·50 0·22 7·61 11·28 2·03 0·06 0·14 99·68 0·80 <1·5 5·6 218 96 389 28 43 105 107 56 1703 73 126

6

TE

JAD

Aet

al.O

RIG

INA

ND

EV

OL

UT

ION

OF

ON

TO

NG

JAV

APL

AT

EA

U


KF-30 49·56 1·52 13·95 13·81 0·28 7·10 11·12 1·83 0·12 0·12 99·41 0·57 <1·5 4·8 135 88 369 27 41 101 115 50 2326 66 119

KF-31 49·74 1·57 14·20 13·93 0·20 6·75 10·95 2·04 0·14 0·13 99·65 0·74 1·8 5·1 146 93 334 30 33 94 111 45 1325 66 99

KF-32 49·65 1·27 15·87 12·06 0·18 7·04 12·25 1·85 0·09 0·11 100·37 0·86 <1·5 4·0 151 74 265 24 33 72 120 45 1274 91 154

KF-36 49·60 1·62 14·08 13·78 0·20 6·74 11·45 1·98 0·14 0·12 99·71 0·90 1·7 5·0 143 92 343 30 36 121 108 52 1463 75 120

KF-43 49·84 1·66 14·25 14·06 0·20 6·69 10·99 2·08 0·12 0·14 100·03 0·55 <1·5 5·3 145 98 337 30 35 106 105 47 1397 59 95

KF-49 49·25 1·70 14·54 12·22 0·21 6·81 12·57 1·93 0·15 0·13 99·51 2·23 1·5 5·5 148 91 401 32 43 90 111 52 1525 63 108

KF-53 49·89 1·56 14·70 13·19 0·19 6·73 11·20 2·06 0·11 0·13 99·76 0·41 <1·5 5·4 139 91 395 30 49 106 113 53 1537 81 170

University of Hawaii analyses of standards

BCR-1 54·39 2·36 13·60 13·65 0·18 3·58 6·98 3·52 1·71 0·39 100·35 47·4 12 329 195 410 37 31 126 18 36 1331 12 5

SD (n = 6) 0·08 0·05 0·02 0·05 0·00 0·02 0·03 0·04 0·01 0·00 0·1 0·2 1 1 3 0 1 1 1 2 10 1 3

BHVO-1 49·47 2·85 13·69 12·41 0·17 7·24 11·36 2·43 0·53 0·29 100·42 8·8 18·4 387 177 317 27 32 100 133 45 1280 115 284

SD (n = 6) 0·09 0·06 0·04 0·02 0·00 0·01 0·02 0·03 0·01 0·00 0·1 0·2 1 0 3 0 1 2 1 2 2 2 0

Recommended values (Govindaraju, 1994)

BCR-1 54·44 2·25 13·72 13·54 0·18 3·50 6·99 3·29 1·70 0·36 99·98 47·2 14 330 190 407 38 33 130 19 37 1371 13 16

BHVO-1 49·59 2·69 13·70 12·27 0·17 7·18 11·32 2·24 0·52 0·27 99·95 9·5 19 390 179 317 28 32 105 136 45 1262 121 289

Major element oxides are in weight percent; trace elements are in ppm. LOI, weight loss on ignition to 900°C for at least 10 h. Fe2O3∗, all Fe as Fe2O3. Estimatedrelative uncertainties on major and minor elements are 1%. For trace elements, estimated average overall uncertainties based on measurements of standardsAGV-1, BCR-1, BHVO-1, and W-1 are (in ppm): 0·3 for Rb, 0·2 for Nb, 1·0 for Sr, 1·0 for Zr, 3·5 for V, 0·3 for Y, 1·4 for Sc, 2·5 for Cr, 2·1 for Zn, 1·5 for Ni, 2·0 forCu, 2·1 for Co, and 10 for Mn. An indication of accuracy is given by measured values of standards BCR-1 and BHVO-1 and recommended working values(Govindaraju, 1994). The dashed lines define the boundary between the two chemical groups of lavas along the Singgalo and Kwaimbaita river sections.

7

JOU

RN

AL

OF

PET

RO

LO

GY

VO

LU

ME

43N

UM

BE

R3

MA

RC

H2002

Table 2: ICP-MS data for central Malaita basement rocks

Li Cs Rb Ba Th U Nb Ta La Ce Pr Sr Nd Zr Hf Sm Eu Gd Tb Dy Y Ho Er Tm Yb Lu Mo Ga Co Cr

Kwaio Anticline

Singgalo River

SG1 6·03 0·01 2·4 28 0·61 0·13 4·9 0·35 4·75 12·6 1·98 138 9·81 81 2·50 3·23 1·13 4·14 0·73 4·56 25 1·00 2·90 0·41 2·83 0·43 0·59 17 51 56

SG2 10·41 0·01 2·4 30 0·57 0·13 4·9 0·34 4·62 12·5 1·83 145 9·48 79 2·40 3·00 1·17 3·79 0·68 4·21 23 0·93 2·58 0·36 2·51 0·38 0·49 16 46 81

SG3 8·17 0·02 2·5 33 0·61 0·14 5·6 0·37 5·15 14·0 2·10 146 10·6 89 2·88 3·22 1·10 4·10 0·73 4·45 26 1·00 2·93 0·41 2·81 0·42 0·53 22 61 114

SG4 7·02 0·03 0·9 31 0·44 0·16 5·6 0·39 5·50 14·4 2·32 144 11·6 94 2·79 3·43 1·21 4·44 0·79 4·91 28 1·08 3·12 0·44 ·05 0·46 0·66 19 51 74

SG5 5·60 0·03 2·7 38 0·52 0·15 5·4 0·37 5·42 14·3 2·17 152 10·9 91 2·64 3·43 1·27 4·37 0·75 4·91 25 0·99 2·79 0·42 3·12 0·49 0·36 15 39 62

SG6 6·56 0·02 1·4 31 0·45 0·13 5·1 0·36 5·33 13·8 1·93 148 9·80 83 2·61 3·12 1·09 4·02 0·72 4·57 26 0·98 2·84 0·40 2·75 0·42 0·49 17 48 95

SG7 5·97 0·01 1·0 33 0·60 0·16 5·9 0·40 5·90 15·9 2·27 148 11·6 96 2·83 3·69 1·23 4·70 0·84 5·14 29 1·11 3·10 0·44 3·00 0·45 1·13 19 50 84

SGB1 6·46 0·01 2·4 37 0·59 0·15 5·9 0·40 5·63 14·8 2·15 148 11·3 93 2·97 3·41 1·20 4·38 0·75 4·74 28 1·08 3·06 0·44 3·02 0·46 0·75 19 49 92

SGB2 7·35 0·02 2·8 44 0·54 0·17 6·0 0·41 5·57 14·8 2·26 149 11·0 94 3·03 3·45 1·19 4·25 0·78 4·79 28 1·03 3·00 0·42 2·93 0·45 0·64 19 49 95

SGB6 6·23 0·03 1·6 33 0·49 0·16 5·8 0·38 5·59 15·4 2·21 152 11·2 101 2·89 3·60 1·22 4·56 0·79 5·04 27 1·07 2·98 0·42 2·86 0·43 0·59 18 50 89

SGB7 5·40 0·02 2·2 36 0·56 0·14 5·3 0·36 5·71 15·0 2·46 141 12·5 93 2·75 3·54 1·28 4·44 0·85 5·17 29 1·07 3·16 0·47 3·10 0·47 0·61 18 48 91

SGB8 6·15 0·02 1·5 37 0·66 0·15 5·8 0·38 5·45 15·3 2·19 146 11·2 99 2·85 3·57 1·23 4·51 0·82 5·08 28 1·09 3·05 0·43 2·91 0·44 1·00 19 47 89

SGB3 6·81 0·02 1·1 30 0·65 0·17 5·9 0·41 5·75 15·3 2·40 159 12·1 95 2·96 3·60 1·25 4·51 0·81 4·95 30 1·07 3·06 0·44 3·02 0·46 0·60 18 49 89

SGB4 6·18 0·02 1·5 30 0·56 0·19 5·7 0·39 5·64 15·0 2·24 146 11·5 94 2·94 3·52 1·25 4·42 0·81 5·03 30 1·07 2·98 0·42 2·91 0·44 0·66 18 48 84

SGB5 6·95 0·02 1·8 29 0·53 0·15 5·7 0·38 5·14 14·4 2·03 141 10·8 95 2·75 3·34 1·15 4·13 0·78 4·80 29 1·02 2·92 0·41 2·84 0·43 0·70 18 47 84

SGB9 4·86 0·01 2·0 32 0·56 0·13 5·1 0·33 6·03 16·0 2·44 144 11·9 77 2·56 3·08 1·08 4·67 0·74 4·38 25 0·94 2·72 0·38 2·64 0·40 0·54 18 48 138

SGB12 6·38 0·01 0·8 33 0·60 0·16 6·4 0·41 5·61 15·8 2·28 162 11·9 99 3·11 3·62 1·32 4·62 0·83 5·29 30 1·12 3·17 0·44 3·02 0·45 0·45 19 49 92

SGB10 5·78 0·02 2·2 36 0·57 0·15 5·8 0·38 5·70 15·7 2·27 142 11·6 94 2·87 3·52 1·24 4·55 0·82 5·12 29 1·08 3·09 0·43 2·98 0·45 0·86 18 47 83

SGB11 5·37 0·02 0·5 25 0·64 0·15 5·9 0·38 5·71 15·5 2·22 150 11·4 100 2·84 3·54 1·21 4·53 0·82 5·05 30 1·08 3·10 0·43 2·97 0·45 2·26 18 49 83

SGB13 6·52 0·01 1·5 37 0·52 0·15 6·2 0·40 6·26 16·8 2·38 155 12·2 101 2·93 3·89 1·31 4·83 0·86 5·34 30 1·14 3·22 0·45 3·05 0·46 0·63 18 49 81

SGB14 6·51 0·02 2·2 36 0·61 0·16 5·9 0·39 5·64 15·4 2·16 146 11·0 99 2·91 3·54 1·21 4·51 0·79 4·99 29 1·08 3·01 0·42 2·89 0·44 0·86 19 49 79

SGB16 5·50 0·02 1·0 31 0·62 0·14 5·7 0·35 5·53 15·1 2·30 140 11·6 97 2·70 3·33 1·23 4·51 0·80 4·77 28 0·99 2·80 0·42 2·93 0·44 0·57 18 50 100

SGB15 6·38 0·01 1·4 17 0·39 0·11 4·5 0·30 4·38 12·1 1·72 136 8·82 78 2·20 2·88 1·02 3·72 0·69 4·11 24 0·92 2·49 0·35 2·33 0·35 0·46 17 51 103

SGB17 6·07 0·02 1·9 20 0·50 0·13 5·0 0·32 4·69 12·9 1·86 134 9·69 83 2·47 3·05 1·04 3·81 0·70 4·37 27 0·96 2·64 0·37 2·51 0·38 0·53 17 50 88

SGB18 4·91 0·01 1·4 13 0·35 0·08 3·7 0·23 3·30 9·75 1·41 125 7·64 69 2·00 2·48 0·88 3·20 0·60 3·76 21 0·83 2·29 0·32 2·16 0·32 0·37 16 50 133

SGB22 4·17 0·01 1·0 13 0·34 0·09 3·9 0·26 3·41 9·77 1·48 127 7·76 68 2·08 2·55 0·92 3·34 0·62 3·86 23 0·85 2·34 0·33 2·21 0·33 0·82 16 53 133

SGB23 4·71 1·2 14 0·34 0·10 4·1 0·27 3·56 10·2 1·50 123 7·97 71 2·06 2·64 0·93 3·44 0·64 4·01 21 0·89 2·49 0·35 2·35 0·35 0·74 16 52 123

SGB24 7·72 0·02 2·5 41 0·28 0·07 3·2 0·21 3·15 9·29 1·34 211 7·07 52 1·42 2·39 0·89 3·06 0·57 3·51 16 0·74 2·10 0·30 2·05 0·31 0·27 16 47 127

SGB25 4·60 0·01 0·46 12 0·34 0·09 3·7 0·25 3·52 9·58 1·51 126 8·12 68 1·98 2·55 0·91 3·46 0·59 3·67 21 0·81 2·23 0·34 2·37 0·36 0·18 16 53 139

8

TE

JAD

Aet

al.O

RIG

INA

ND

EV

OL

UT

ION

OF

ON

TO

NG

JAV

APL

AT

EA

U

Li Cs Rb Ba Th U Nb a La Ce Pr Sr Nd Zr Hf Sm Eu Gd Tb Dy Y Ho Er Tm Yb Lu Mo Ga Co Cr

Kwaimbaita River

ML407 4·99 0·01 2·9 23 0·33 0·09 5·0 0·25 3·86 10·6 1·70 288 8·75 83 2·63 2·97 0·94 3·54 0·68 4·22 23 0·92 2·57 0·36 2·47 0·37 0·36 17 50 61

ML476 5·20 0·02 1·4 13 0·56 0·13 6·5 0·40 5·44 16·4 2·26 129 11·9 102 2·95 3·73 1·28 4·71 0·86 5·25 32 1·10 3·19 0·44 3·01 0·45 0·39 22 52 101

ML475 3·51 0·01 0·9 11 0·10 0·04 2·7 0·16 1·61 4·80 0·72 135 3·78 30 1·09 1·30 0·51 1·74 0·32 2·05 10 0·43 1·26 0·17 1·16 0·17 0·18 13 39 446

ML474 7·00 0·01 0·3 12 0·36 0·12 4·7 0·27 3·77 10·4 1·58 160 8·49 74 2·80 2·69 1·02 3·44 0·64 4·21 26 0·87 2·58 0·38 2·59 0·42 0·15 17 46 122

ML436 8·59 0·01 0·8 20 0·31 0·09 4·0 0·25 3·48 10·6 1·50 140 8·04 73 1·98 2·62 0·89 3·29 0·59 3·65 22 0·76 2·16 0·31 2·19 0·34 0·21 17 51 155

ML434 9·76 1·3 28 0·28 0·09 3·8 0·23 3·36 10·1 1·44 147 7·64 69 2·23 2·49 0·85 3·09 0·56 3·40 23 0·70 2·06 0·29 1·94 0·29 0·27 16 47 143

ML432 8·79 0·01 1·6 30 0·31 0·09 3·8 0·24 3·00 9·10 1·34 169 7·30 65 2·00 2·43 0·89 3·04 0·57 3·45 19 0·74 2·09 0·29 1·98 0·30 0·30 17 54 130

ML431 4·78 0·01 0·6 15 0·34 0·08 3·6 0·24 3·49 10·3 1·43 121 7·69 67 1·91 2·40 0·90 3·14 0·58 3·64 22 0·76 2·20 0·31 2·10 0·32 0·85 16 51 128

ML423 4·17 0·5 12 0·27 0·08 3·1 0·19 3·44 9·84 1·51 130 7·93 66 2·12 2·53 0·89 3·31 0·60 3·72 16 0·82 2·38 0·33 2·29 0·35 0·75 19 51 128

ML422 4·22 0·7 14 0·25 0·08 3·4 0·22 3·30 9·60 1·46 157 8·09 64 2·29 2·63 0·96 3·42 0·61 3·81 19 0·86 2·43 0·34 2·34 0·35 0·52 13 42 100

ML451 20·4 0·01 0·8 16 0·32 0·07 2·9 0·20 2·95 8·87 1·25 115 6·85 56 1·63 2·25 0·82 2·86 0·55 3·31 21 0·69 1·98 0·28 1·88 0·28 0·33 16 49 134

ML450 5·20 2·5 31 0·55 0·13 5·2 0·34 5·18 13·6 2·13 140 10·4 82 2·50 3·21 1·22 4·21 0·74 4·56 24 0·95 2·76 0·39 2·59 0·44 0·49 19 46 86

ML449 8·05 0·01 2·1 39 0·44 0·16 6·1 0·39 5·72 16·1 2·19 146 11·1 99 2·88 3·45 1·20 4·32 0·76 4·62 29 0·98 2·75 0·38 2·62 0·39 0·45 21 50 100

ML421 5·96 0·01 0·4 21 0·47 0·16 6·2 0·40 5·46 14·9 2·21 152 11·0 99 2·94 3·56 1·22 4·28 0·76 4·81 27 1·05 2·96 0·42 2·84 0·43 0·46 19 49 91

ML437 4·55 0·02 2·0 42 0·58 0·11 4·7 0·29 4·80 13·6 1·83 124 9·65 86 2·17 3·03 1·06 3·89 0·69 4·32 27 0·91 2·60 0·36 2·48 0·37 1·17 19 50 110

ML438 6·83 0·02 2·0 31 0·58 0·14 5·7 0·35 5·33 15·0 2·08 135 10·7 94 2·58 3·33 1·15 4·19 0·73 4·47 27 0·95 2·69 0·37 2·54 0·38 0·50 19 50 74

ML439 8·82 0·02 0·5 19 0·58 0·16 6·2 0·40 5·76 16·5 2·24 153 11·8 99 2·94 3·55 1·21 4·39 0·77 4·74 28 1·01 2·85 0·40 2·72 0·41 0·59 19 48 65

ML440 10·11 0·02 2·0 27 0·45 0·13 4·9 0·31 4·61 13·1 1·78 130 9·36 79 2·35 2·85 1·02 3·69 0·64 4·01 26 0·85 2·43 0·34 2·30 0·34 1·04 18 48 108

ML448 6·32 0·01 1·8 33 0·44 0·11 6·0 0·37 5·61 16·2 2·23 143 11·4 76 2·25 3·43 1·17 4·40 0·77 4·73 31 1·01 2·68 0·38 2·56 0·38 0·54 20 51 95

ML468 6·88 0·02 0·6 41 0·67 0·16 6·2 0·42 4·98 14·5 2·02 156 10·5 98 3·02 3·31 1·16 4·03 0·72 4·51 28 0·94 2·65 0·37 2·55 0·38 0·75 20 46 71

9

JOU

RN

AL

OF

PET

RO

LO

GY

VO

LU

ME

43N

UM

BE

R3

MA

RC

H2002

Table 2: continued

Li Cs Rb Ba Th U Nb Ta La Ce Pr Sr Nd Zr Hf Sm Eu Gd Tb Dy Y Ho Er Tm Yb Lu Mo Ga Co Cr


Kwaiafa’a River

KF1 6·48 0·11 3·3 133 0·27 0·08 3·5 0·24 4·06 11·0 1·64 116 8·14 60 1·73 2·55 1·08 3·14 0·55 3·49 18 0·71 2·11 0·29 1·94 0·29 0·17 18 55 155

KF4 5·27 0·04 2·3 28 0·56 0·15 6·0 0·39 6·38 19·6 2·94 137 14·0 106 3·02 4·25 1·65 5·17 0·90 5·84 30 1·18 3·49 0·50 3·30 0·47 0·47 22 51 34

KF9 6·67 0·02 2·5 32 0·52 0·15 5·6 0·38 6·11 18·2 2·72 142 13·4 100 2·80 3·97 1·54 4·84 0·87 5·57 29 1·11 3·29 0·48 3·15 0·45 0·80 22 55 113

KF13 6·34 0·01 0·6 16 0·68 0·19 6·1 0·39 6·27 19·1 2·84 145 14·1 125 3·45 4·29 1·59 5·12 0·92 6·04 32 1·23 3·69 0·53 3·51 0·51 0·31 22 56 101

KF14 5·73 0·01 0·8 24 0·53 0·15 5·5 0·37 5·96 18·0 2·63 145 13·0 98 2·78 3·92 1·48 4·79 0·86 5·46 28 1·12 3·21 0·46 3·08 0·44 0·96 21 54 100

KF19 5·51 0·01 0·9 24 0·45 0·15 5·2 0·34 4·97 15·4 2·34 141 11·7 93 2·65 3·54 1·40 4·37 0·77 5·00 24 1·01 2·98 0·41 2·83 0·40 2·76 21 53 99

KF24 7·63 0·04 0·7 50 0·56 0·17 5·8 0·37 5·76 17·7 2·69 207 13·0 106 2·99 3·87 1·54 4·91 0·85 5·56 28 1·12 3·23 0·46 3·12 0·44 0·42 21 60 94

KF30 7·26 0·01 2·0 26 0·64 0·16 5·7 0·37 5·83 17·9 2·70 134 13·4 106 2·98 4·02 1·56 4·93 0·86 5·63 29 1·14 3·38 0·47 3·38 0·47 1·49 21 55 97

KF31 5·31 0·01 1·9 30 0·53 0·15 5·5 0·36 6·06 18·1 2·70 134 13·2 98 2·82 4·09 1·52 4·90 0·84 5·51 29 1·13 3·31 0·46 3·07 0·45 0·61 21 55 99

KF32 4·84 0·01 1·6 22 0·38 0·11 4·1 0·27 4·43 13·4 2·03 138 9·77 73 2·10 2·98 1·18 3·66 0·65 4·24 22 0·86 2·55 0·36 2·42 0·34 0·31 20 55 168

KF36 5·39 0·01 2·5 28 0·53 0·15 5·4 0·36 5·76 17·0 2·52 140 12·8 94 2·70 3·90 1·49 4·69 0·83 5·41 28 1·12 3·26 0·47 3·06 0·44 0·58 21 54 105

KF43 5·77 0·01 1·6 29 0·57 0·16 5·9 0·38 6·07 18·2 2·72 143 13·5 104 2·98 3·99 1·54 4·83 0·88 5·59 28 1·15 3·36 0·48 3·22 0·46 0·97 21 52 79

KF49 7·61 0·02 2·0 37 0·53 0·15 5·5 0·36 6·21 18·5 2·75 145 13·4 96 2·76 4·06 1·55 5·06 0·90 6·03 31 1·20 3·60 0·52 3·39 0·48 0·32 21 55 91

KF53 7·07 0·02 2·0 57 0·51 0·14 5·4 0·35 5·79 17·3 2·59 142 12·9 94 2·78 3·96 1·49 4·90 0·86 5·60 29 1·14 3·39 0·48 3·14 0·46 0·44 21 56 158

University of Notre Dame analyses of standard reference materials

BHVO-1 4·52 0·11 9·4 135 1·20 0·41 19 1·23 15·6 39·8 5·52 409 24·7 170 4·48 6·13 2·11 6·19 0·95 5·37 24 0·96 2·53 0·32 2·07 0·26 1·01 22 47 306

2� (n = 52) 0·38 0·04 0·4 2 0·07 0·04 2·0 0·12 0·25 1·34 0·18 14 0·48 6 0·25 0·17 0·08 0·38 0·06 0·11 1 0·06 0·12 0·04 0·08 0·04 0·21 1 2 29

Recommended values (Govindaraju, 1994)

BHVO-1 4·60 0·13 11 139 1·08 0·42 19 1·23 15·8 39 5·7 403 25·2 179 4·38 6·2 2·06 6·4 0·96 5·2 28 0·99 2·4 0·33 2·02 0·29 1·02 45 289

All abundances are in ppm. Accuracy is illustrated by the published and measured values of BHVO-1. Estimated uncertainties from 52 analyses of BHVO-1standard are <10% for most elements except for Ba, La, Nd, Dy (<5%), Th, Nb, Gd, Tb, Ho (<15%), U, Ta, Cr (19%), Tm (27%), and Lu (30%). Recommended valuesare from Govindaraju (1994). Errors on the means for the BHVO-1 data are 2 SD. The dashed lines define the boundary between the two chemical groups of lavasalong the Singgalo and Kwaimbaita river sections.

10

TE

JAD

Aet

al.O

RIG

INA

ND

EV

OL

UT

ION

OF

ON

TO

NG

JAV

APL

AT

EA

U

Table 3: 40Ar–39Ar incremental heating ages for basalts from central Malaita, Solomon Islands

Sample no. and Depth (m) Total fusion Plateau age 39Ar Isochron age N ( 40Ar/36Ar) J

location age (Ma) (Ma) % of total (Ma) Intercept ± 1�

Kwaio Anticline

Singgalo River

SG7 172 124·2 123·1 ± 2·1 100·0 118·4 ± 14·2 4 327 ± 105 0·001463

SGB10 373 126·8 126·1 ± 8·5 43·6 125·6 ± 4·6 5 336 ± 143 0·001397

SGB25 889 117·0 128·2 ± 8·5 62·5 126·6 ± 17·9 3 297 ± 11 0·001529

Kwaimbaita River

ML423 893 104·5 none developed

ML475 2743 119·6 123·2 ± 3·0 90·4 125·1 ± 8·3 5 289 ± 34 0·001582


Kwaiafa’a River

KF14 315 121·6 120·0 ± 2·2 46·6 118·9 ± 4·5 4 299 ± 16 0·001397

KF36 371 127·2 120·7 ± 3·2 61·8 120·7 ± 30·6 4 313 ± 434 0·001747

KF53 31 124·6 124·9 ± 3·8 58·7 149·6 ± 48·8 5 −104 ± 1334 0·001774

Ages are reported relative to biotite monitor FCT-3 (27·55 ± 0·12 Ma), which is calibrated against hornblende Mmhb-1 (513·9 Ma, Lanphere et al., 1990). Plateauages are the mean of concordant step ages (N, number of steps), weighted by the inverse of their variances. Calculations use the following decay and reactorinterference constants: �� = 0·581 × 10−10 yr−1, �� = 4·963 × 10−10 yr−1; ( 36Ar/37Ar)Ca = 0·64, (39Ar/37Ar)Ca = 0·000673, (40Ar/39Ar)Ca = 0·01. J is the neutron fluencefactor, determined from measured monitor40Ar/39Ar.

11


Sm, Pb, Th, and U, used for isotopic age corrections, occurring at locations separated by as much as 1600 km(see Tejada et al., 1996). Furthermore, the combined dataare presented together with isotopic ratios in Table

4. As in our previous studies, elemental abundances now available suggest that the >90 Ma event is notrepresented on Malaita.determined by isotope dilution do not necessarily rep-

resent those of the bulk rocks because, for isotopic work,small chips were handpicked from the bulk rock andcleaned in acid before powdering; also, some meas- Volcanic stratigraphyurements were made on strongly acid-leached splits of

Effects of alteration on chemical compositionthese powders. In addition to the basaltic lavas, a plagio-Relative to many of the samples from southern Malaitaclase megacryst and its host basalt (SGB21) from a sampleand Santa Isabel studied by Tejada et al. (1996), theof the ‘orbicular’ facies rocks were analyzed isotopicallymajority of the central Malaitan lavas are only weaklyto evaluate the genetic relationship of the megacrysts andaltered, as indicated by, for example, their relatively lowlavas.(<1·5 wt %) weight loss on ignition (LOI) values (Table1); also, only a few samples with MgO values between7·0 and 8·5 wt % have comparatively high K2O andNa2O values (>0·25 and >3·00 wt %, respectively). ThisRESULTSis consistent with Babbs’ (1997) conclusion that alteration40Ar–39Ar agesis limited mainly to clays replacing olivine and mesostasis.Previous work on the OJP has revealed two distinct setsAs a result, data for elements and oxides known to beof basement ages (Fig. 3). Lavas in southern Malaita,affected by higher levels of submarine alteration, such asRamos Island, parts of Santa Isabel, at DSDP Site 289,Na2O, CaO, P2O5, Cr, Ni, Zn, Cu, and V (e.g. Bienvenuand at ODP Site 807 all have ages that are in-et al., 1990), still show some coherence with data fordistinguishable, within errors, around a mean of 122 Mamore immobile elements such as TiO2, Fe2O3, Al2O3,(total range is 119–125 Ma). In contrast, basalts at ODPNb, Y, Zr, and the lanthanide rare earths. Nevertheless,Site 803, in much of Santa Isabel and parts of Sanmost of our petrogenetic interpretations are based on theCristobal, and ash layers at DSDP Site 288 yield agesalteration-resistant elements.around 90 Ma (total range is 86–94 Ma; Mahoney et al.,

1993; Birkhold-VanDyke et al., 1996; Parkinson et al.,Isotopic and chemical variations1996; Tejada et al., 1996). Compositionally similar thole-

iites with 40Ar–39Ar ages of >62 Ma and >34 Ma are Kwaio Anticline. Previous studies have revealed two distinct,ocean-island-like isotopic groups of OJP basement lavas;also present on San Cristobal (Birkhold-VanDyke et al.,

1996), but their relationship to the older OJP basalts is at Site 807, where both groups were sampled, they arestratigraphically separated (Mahoney & Spencer, 1991;as yet unclear. Our 40Ar–39Ar ages for rocks from the

Singgalo, Kwaimbaita, and Kwaiafa’a river sections all Mahoney et al., 1993; Tejada et al., 1996). Two strati-graphically separated, isotopically distinct groups of lavasfall within the >122 Ma group (Table 3). Five of the

samples produced acceptable plateau ages; that is, con- also are present in the Kwaio Anticline (Fig. 4; Table4). The upper stratigraphic group has initial �Nd(t) =cordant step-ages comprising >50% of the total 39Ar

released. An additional two samples (SGB10, KF14) gave +3·8 to +3·9, ( 87Sr/86Sr)t = 0·7040–0·7042, ( 206Pb/204Pb)t = 17·71–17·85, ( 207Pb/204Pb)t = 15·47–15·49,similar plateau ages, based on slightly less than half of

the total gas; one sample (ML423) produced a recoil and ( 208Pb/204Pb)t = 37·87–37·98. The stratigraphicallylower group has �Nd(t) = +5·4 to +6·0, ( 87Sr/86Sr)t =age-spectrum pattern, from which only a minimum (in-

tegrated) age estimate could be determined. Isochron 0·7037–0·7039, ( 206Pb/204Pb)t = 18·12–18·40, ( 207Pb/204Pb)t= 15·51–15·53, and ( 208Pb/204Pb)t= 38·15–38·23ages are concordant with plateau ages for all samples,

but generally have larger uncertainties because of the (Figs 4–6). Basalts of the Kwaelungga River sectionsouth of the Singgalo River (Babbs, 1997) show similarsmall dispersion of the few step compositions. For this

reason, we take the acceptable plateau ages as the best characteristics. The plagioclase megacryst of sampleSGB21 has an identical isotopic composition, withinestimate of the timing of volcanism. These five ages are

indistinguishable at the 95% confidence level, and provide error, to that of its host rock, which belongs with thelower stratigraphic group. Malaita Volcanic Group lavasa mean age for the entire section of 123·1 ± 1·4 Ma.

There is no evidence of any significant time difference in southern Malaita (Mahoney & Spencer, 1991; Tejadaet al., 1996) are isotopically equivalent to the upper group.between the lowermost and uppermost lavas, from either

the 40Ar–39Ar ages or the volcanic stratigraphy (e.g. Moreover, the two isotopic groups in the Kwaio Anticlineare virtually indistinguishable from those of northeasternsedimentary interbeds are almost nonexistent). Combined

with previously reported ages, these results show the Santa Isabel (Tejada et al., 1996) and similar to the upper(Unit A) and lower (Units C–G) groups, respectively, of>122 Ma volcanic event was widespread across the OJP,

12

TE

JAD

Aet

al.O

RIG

INA

ND

EV

OL

UT

ION

OF

ON

TO

NG

JAV

APL

AT

EA

U

Table 4: Isotope ratios and isotope dilution abundances for central Malaita, Solomon Islands

Sample (87Sr/86Sr)t �Nd(t) ( 206Pb/ (207Pb/ (208Pb/ Rb Sr Sm Nd (206Pb/ (207Pb/ (208Pb/ Th U Pb204Pb)t

204Pb)t204Pb)t

204Pb)0204Pb)0

204Pb)0

Kwaio Anticline

Singgalo River

SG-1 U 0·70404 +3·9 17·847 15·472 37·913 1·277 107·8 1·689 4·607 18·511 15·505 38·537 0·2772 0·0966 0·1710

SGB13 U 0·70417 +3·7 1·471 147·6 3·650 12·02

L 0·70415 +3·9 0·9029 121·5 1·480 3·293

SGB17 U 17·708 15·488 37·865 18·373 15·520 38·423 0·3003 0·1174 0·2067

L 0·70413 +3·9 0·9675 93·13 1·540 3·425

SGB18 U 18·117 15·513 38·215 18·570 15·535 38·633 0·2217 0·1040 0·2048

L 0·70378 +5·7 0·6878 112·5 1·439 3·253

SGB21 P 0·70372 +5·4 0·0778 187·4 0·0388 0·1767 18·369 15·523 38·180 0·0273

SGB21 U 0·70379 +5·8 18·403 15·514 38·199 1·311 67·44 1·591 4·150 18·963 15·541 38·554 0·1290 0·0667 0·1409

SGB25 U 0·70382 +6·0 18·245 15·506 38·147 0·1820 133·3 2·116 5·993 18·466 15·517 38·351 0·1526 0·0541 0·2869

Kwaimbaita River

ML407 U 0·70413 +5·7 18·322 15·511 38·230 1·769 100·1 2·110 5·672 18·694 15·529 38·635 0·1553 0·0467 0·1483

L 0·70393 +5·7 0·8105 46·78 1·981 4·490

ML475 U 0·70371 +5·6 18·352 15·526 38·178 0·3530 74·86 0·9217 2·281 18·568 15·537 38·405 0·0705 0·0219 0·1195

ML422 L 0·70378 +5·8 0·1206 94·20 1·510 3·357

ML451 L 0·70378 +5·5 0·2478 43·10 1·264 2·633

ML450 L 0·70413 +3·8 0·4012 76·73 1·358 2·968

ML421 U 0·70423 +3·8 17·756 15·468 37·984 0·3208 147·3 3·488 11·34 18·156 15·488 38·206 0·2140 0·1267 0·3687

L 0·70415 +3·9 0·0990 109·6 1·326 2·885


Kwaiafa’a River

KF1 L 0·70441 +3·9 1·360 62·23 1·281 2·928

D 0·70446 1·210 59·28

KF4 U 0·70414 +3·9 17·909 15·490 38·028 1·120 115·7 1·653 4·530 18·301 15·509 38·380 0·1921 0·0700 0·2090

L 0·70415 +3·8 0·7995 81·33 1·558 3·673

L, strongly acid-leached powder; U, unleached but chips handpicked and acid-cleaned; P, plagioclase; D, duplicate analysis. Isotopic fractionation corrections are148NdO/144NdO = 0·242436 (148Nd/144Nd = 0·241572), 86Sr/88Sr = 0·1194. Data are reported relative to University of Hawaii standard values: for La Jolla Nd, 143Nd/144Nd = 0·511842; for NBS 987 Sr, 87Sr/86Sr = 0·71024. The total range measured for La Jolla Nd is ±0·000008 (0·2� units); for NBS 987 it is ±0·000020 over a1 yr period. Present-day Pb isotopic ratios are corrected for fractionation using the NBS 981 standard values of Todt et al. (1996); the total ranges measured forNBS 981 are ±0·012 for 206Pb/204Pb, ±0·012 for 207Pb/204Pb, and ±0·038 for 208Pb/204Pb. Within-run errors on the isotopic data above are less than or equal to theexternal uncertainties on these standards. Estimated uncertainties on abundances are <0·2% on Sm and Nd; <0·5% for Sr, % for Rb; <2% on Th, <1% on U, and<1% on Pb. Total blanks are negligible: <3 pg for Th, <5 pg for U, and 5–30 pg for Pb. All abundances are in ppm. Total blanks are negligible: <15 pg for Nd and<60 pg for Sr. �Nd = 0 today corresponds to 143Nd/144Nd = 0·51264; for 147Sm/144Nd = 0·1967, �Nd(t) = 0 corresponds to 143Nd/144Nd = 0·512486 at 120 Ma.

13


3·4± 0·5 ppm, Ce 14·4± 2·4 ppm vs 9·8± 0·8 ppm,Nb 5·9 ± 0·5 ppm vs 4·0 ± 1·0 ppm, and Th 0·53 ±0·14 ppm vs 0·30 ± 0·06 ppm. Except for samplesML407, ML475, and ML476, Cr and Ni contents aregenerally higher in the lower group of lavas (128 ± 28vs 112 ± 56 ppm and 110 ± 15 vs 68 ± 28 ppm,respectively), probably indicating slightly lesser amountsof crystal fractionation than in the stratigraphically uppergroup. Sample ML475, with 446 ppm Cr, appears tocontain abundant excess (cumulus) clinopyroxene

Kwara’ae Anticline. The basalt flows exposed in theKwara’ae Anticline are chemically very similar to thestratigraphically upper group of lavas in the KwaioAnticline; for example, in TiO2 (1·5 ± 0·2 wt %), P2O5

(0·11 ± 0·03 wt %), La (5·2 ± 1·2 ppm), Ce (15·3 ±4·3 ppm), Nb (4·8 ± 1·3 ppm), and Th (0·48 ± 0·20

Fig. 3. Summary of 40Ar–39Ar and biostratigraphic (points with error ppm; Tables 1 and 2). The two samples analyzed iso-bars) age data for OJP tholeiites, ash layers at DSDP Site 288, and topically, KF1 and KF4, from the top and bottom,late-stage alkalic rocks. Data sources are this paper, Davis (1978),

respectively, of the Kwaiafa’a River section (Figs 4–6)Berger et al. (1993), Mahoney et al. (1993), Birkhold-VanDyke et al.(1996) and Tejada et al. (1996). Φ, mean ages reported by Parkinson also have similar isotopic compositions to the upperet al. (1996). group of Kwaio Anticline lavas, except that KF1 has

higher ( 87Sr/86Sr)t (0·70441) even after multi-step acidleaching. This sample was taken from the topmost lava

lavas at ODP Site 807 (Mahoney et al., 1993). In detail, flow, in contact with the Kwara’ae Mudstone and sep-the upper group in the Kwaio Anticline has slightly lower arated from the rest of the lava sequence by a thin bed�Nd(t) and Pb isotope ratios than the flows of Unit A at of chert. Its high ( 87Sr/86Sr)t value is probably a result ofSite 807 (e.g. with �Nd(t) = +4·5 to +5·1), whereas the seawater alteration.lower group possesses slightly higher ( 87Sr/86Sr)t than theUnits C–G basalts (with values of 0·7034–0·7035), much

Proposed formation names, estimated thickness, and regionallike the lavas of northeastern Santa Isabel (0·7036–correlations0·7037).

Except for compositionally extreme samples ML475 We propose that the geochemically distinct upper(MgO 10·0 wt %) and ML476 (MgO 5·8 wt %), the major sequence of flows in central Malaita, characterized byelement compositions of the Kwaio Anticline basalts have lower �Nd(t) and ( 206Pb/204Pb)t, higher ( 87Sr/86Sr)t, anda relatively small range in MgO (6·3–8·9 wt %), molar greater abundances of the more highly incompatiblemg-number [100 × Mg/(Mg + Fe2+) = 51–61, as- elements, be designated as the Singgalo Formation (Fm.),suming Fe2+ is 0·85 of total Fe], SiO2 (48·6–50·5 wt %), and that the lower sequence, with higher �Nd(t) andTiO2 (0·9–1·8 wt %), Al2O3 (13·5–15·9 wt %), CaO ( 206Pb/204Pb)t, and lower ( 87Sr/86Sr)t and incompatible-(10·4–13·2 wt %), and total Fe as Fe2O3 (Fe2O3∗ 11·9– element contents, be termed the Kwaimbaita Fm. These14·6 wt %) (Fig. 7; Table 1). These results are consistent designations are appropriate because they are taken fromwith those of previous studies, which have documented the names of the two major rivers where the thickesta much smaller range of major element variation in exposures of each magma type have been found (Figs 2OJP lavas than seen in, for example, the much smaller and 4). On the basis of the close isotopic and elementalManihiki or Caribbean plateaux (e.g. Mahoney et al., affinity noted above,>122 Ma lavas sampled in southern1993; Tejada et al., 1996; Kerr et al., 1997). In general, Malaita, northernmost Malaita, Ramos Island, and manylavas of the stratigraphically upper isotopic group have of those in northeastern Santa Isabel also can be classifiedlower CaO and MgO/Fe2O3∗ but higher TiO2, and as Singgalo type, whereas a few lavas sampled in north-Fe2O3∗ for a given MgO content than the underlying eastern Santa Isabel are Kwaimbaita type (Figs 5–8).lavas (Fig. 7). Isotopic ratios of the single flow sampled at Site 289

As with the two groups of lavas at Site 807, in- indicate that it is also of the Kwaimbaita type. At Sitecompatible elements in the upper group of the Kwaio 807, the upper (Unit A) and lower (Units C–G) groupsAnticline are slightly more abundant than in the under- of flows likewise correspond, respectively, to the Singgalolying group (Tables 1 and 2): for example, TiO2 1·5 ± and Kwaimbaita types, although physical continuity with0·3 (total range) vs 1·1 ± 0·2 wt %, P2O5 0·14 ± the Malaitan volcano-stratigraphic formations remains

to be demonstrated. In contrast, the >90 Ma lavas of0·02 wt % vs 0·10 ± 0·02 wt %, La 5·4 ± 1·0 ppm vs

14


Fig. 4. Initial Pb, Sr, and Nd isotopic ratios, La content, and 40Ar–39Ar ages vs basement depth for central Malaita. The Site 807 basementsection (e.g. Mahoney et al., 1993) is shown for reference adjacent to the basement stratigraphic column in central Malaita. The filled and opensymbols represent lavas of the two groups we term the Singgalo Fm. and Kwaimbaita Fm., respectively (see text); squares indicate samples fromthe Kwaio Anticline, and diamonds indicate samples from the Kwara’ae Anticline. The solid line in the Site 807 section represents a thinlimestone unit; the bold dashed line in the central Malaitan section represents the boundary between the Singgalo Fm. and Kwaimbaita Fm;the fine dashed line correlates the equivalent boundary at Site 807.

Site 803, Santa Isabel, and San Cristobal, although 807 (i.e. Unit A) is only 46 m (Mahoney et al., 1993; Fig.isotopically similar to the >122 Ma basalts, cannot be 4); furthermore, the Singgalo- and Kwaimbaita-type lavarelated directly to either formation because they are some groups at this site are divided by a pelagic limestone unit30 my younger; however, when discussing petrogenesis, of 0·5 m thickness not seen in Malaita. At Site 289,we refer to them as Kwaimbaita-like or Singgalo-like. Singgalo-type lavas are missing entirely. Thus, the Sing-

Petterson (1995) estimated the maximum stratigraphic galo magma type appears likely to be much less vo-thickness of the OJP basement section exposed in central luminous in the northern part of the plateau than in theMalaita to be >4 km. The precise determination of the south.thickness of the Singgalo Fm. is complicated by the For the Kwaimbaita Fm. only a minimum thickness canpresence of faults in lavas of this formation in the relevant be estimated, because its base is not exposed anywhere inportion of the Kwaimbaita River section and the up- the sections we sampled. From the structure of the Kwaiostream portion of the Singgalo River section (Fig. 2b). Anticline and the location of its fold axis (Petterson et al.,However, in the downstream portion of the Singgalo 1997), we estimate that the exposed thickness of theRiver traverse, an apparently continuous and undisturbed Kwaimbaita Fm. is>2700 m in the eastern limb, wheresection exists, in which the top of the Singgalo Fm. is no major faults were found within this formation, andoverlain conformably by the Kwara’ae Mudstone Fm. >2600 m in the western limb (the latter value includesand the base is marked by a sharp change to Kwaimbaita the 310 m exposed in the lower portion of the SinggaloFm. isotopic and chemical compositions (Fig. 4, Table River; see Fig. 2b). At Site 807, only 102 m of Kwaim-1). The Singgalo Fm.’s thickness here is >750 m. The baita-type basalts (i.e. Units C–G) were penetrated beforeSinggalo Fm. appears to be much thicker (>1290 m) in drilling was terminated.the upstream portion of the Singgalo River section, butthe presence of several low-angle faults in this area (Fig.2b) suggests a repetition of strata, which would explain

Chemical variations within each magma typewhy no evidence was found for a significant increaseThe Malaitan lavas display a broader range of chemicalwith depth in the grade of alteration (Babbs, 1997).composition than the basalts at Site 807, probably largelyThe Singgalo Fm. section along the Kwaiafa’a River isreflecting the much greater thickness of basement sampledundisturbed and>500 m thick; however, the base of thein central Malaita. Collectively, however, data for theformation is not exposed. In contrast, the thickness of

the Singgalo-type flows in the basement section at Site few, but widespread, OJP localities sampled thus far

15


indicate that the lavas of each magma type are com- Fig. 7d). Similarly, Kwaimbaita-type lavas can be groupedpositionally very uniform overall, except for rare high- into high- and low-Ti types: (1) those with TiO2 betweenMgO examples such as ML475 and ML68, and rare 1·0 and 1·5 wt %, representing the majority of centralevolved ones such as ML476 and the High-Ti Sigana Malaitan Kwaimbaita Fm. lavas, the Site 807 Units C–GBasalts of Santa Isabel (see Fig. 7d; Tejada et al., 1996). flows, and the Site 289 basalt; (2) those with TiO2

Singgalo-type basalts can conveniently be divided into Ζ1 wt % in central Malaita (lower dashed outline inthree general chemical types (Fig. 7d): (1) a high-Ti group Fig. 7d). The >90 Ma Site 803 and Santa Isabel thole-represented by the High-Ti Sigana Basalts of Santa iites, which are isotopically Kwaimbaita-like, are alsoIsabel; (2) lavas with TiO2 between 1·5 and 2·0 wt %, chemically similar to the high-Ti Kwaimbaita Fm. lavasrepresented by Unit A at Site 807, most of the 122 Ma (see Tejada et al., 1996).Sigana Basalts, and the majority of central, northern, As with Ti, relatively small variations are evidentand southern Malaitan Singgalo-type lavas; (3) basalts in incompatible trace element abundances within eachwith TiO2 <1·5 wt %, overlapping with Site 803 and some magma type, both among flows in different locations andKwaimbaita Fm. compositions (upper dashed outline in at a given location (Fig. 8). For example, the low-Ti

Singgalo Fm. lavas of the Kwara’ae Anticline, KF1 andKF32, have lower incompatible element abundances thanmost other lavas in the same area (Fig. 8b). Regionally,the average incompatible element contents of the Sing-galo-type basalts in central Malaita are slightly higherthan the averages for southern Malaita and Santa Isabel.In addition, the Kwaio Anticline lavas have similaraverage abundances of highly incompatible elements (Zrto Th in Fig. 8) to those of their Site 807 counterparts,but slightly higher average abundances of moderatelyincompatible elements (Sm to Lu). The Kwara’ae An-ticline lavas have slightly greater average abundances ofmost incompatible trace elements than similar units atSite 807, southern Malaita, and Santa Isabel.

Among Kwaimbaita-type basalts, a few central Ma-laitan low-Ti lavas with the lowest incompatible elementabundances (i.e. with patterns at the lower boundary ofthe shaded band in Fig. 8c) display larger troughs at P,

Fig. 5. Initial �Nd(t) vs ( 87Sr/86Sr)t (a) and ( 206Pb/204Pb)t vs �Nd(t) (b) and( 87Sr/86Sr)t (c) for lavas of central Malaita, together with data forsouthern and northernmost Malaita, northeastern Santa Isabel, drillsites 289, 807 and 803 (Mahoney, 1987; Mahoney & Spencer, 1991;Mahoney et al., 1993; Tejada et al., 1996) and additional data fromcentral Malaita (Babbs, 1997). The fields with bold outlines encompassthe drill-site data. Also shown are fields for modern Pacific MORB(Mahoney et al., 1994, and reference therein), estimated initial fieldsfor the Manihiki Plateau (Mahoney and Spencer, 1991), and estimatedinitial values for glasses from the Nauru Basin (Castillo et al., 1994).The shaded field adjacent to the Pacific MORB field is the estimated120 Ma position of the MORB source mantle, assuming the onlychanges in the source have been radioactive decay and ingrowth.Estimated 120 Ma positions of mantle sources of selected Pacific ocean-island basalts are included: Mangaia Group (Vidal et al., 1984; Palacz& Saunders, 1986; Nakamura & Tatsumoto, 1988; Chauvel et al.,1992), Koolau (Stille et al., 1983; Roden et al., 1994), Mauna Loa(Rhodes & Hart, 1995), Kilauea (Tatsumoto, 1978; Chen et al., 1996;Garcia et al., 1996). The 120 Ma mantle source fields assume thefollowing 87Rb/86Sr, 147Sm/144Nd, and 238U/204Pb values: 0·02, 0·24,and 5 for the MORB source (W. M. White, 1993; Mahoney et al.,1998); 0·015, 0·202, and 18 for the Hawaiian source (Roden et al.,1994; Garcia et al., 1995; Tatsumoto, 1978); and 0·054, 0·20, and 22for the Mangaia Group source (Chauvel et al., 1992). Analytical errorson data in this study are smaller than or similar to the size of symbols.

16


basement lavas reveal a strongly bimodal distribution ofages, at >122 Ma and >90 Ma (Fig. 3). For the olderevent, the total range of 40Ar–39Ar ages (excluding thosefor samples SGB10 and SGB25, which have large un-certainties; Table 3) is 119–125 Ma, yielding a maximumpossible duration of 6 my for this event. Table 5 listsestimates of the average emplacement rates, or magmaticfluxes, corresponding to maximum and minimum OJPvolumes estimated by Coffin & Eldholm (1994) and Nealet al. (1997), the smaller value assuming that the plateauwas emplaced entirely in an off-ridge setting and thelarger assuming, unrealistically, that all emplacementtook place at a mid-ocean ridge. Assuming that the>90± 4 Ma event recorded in lavas at Site 803, Santa Isabel,and San Cristobal and in ash layers at Site 288 wasvolumetrically negligible and that essentially the entireplateau was emplaced in 6 my between 119 and 125 Magives minimum average emplacement rates of4·5–10·2 km3/yr. These values are at the low end ofprevious estimates of 16–20 km3/yr (which assumed aduration of only 3 my; Coffin & Eldholm, 1994) and8–22 km3/yr (assuming a duration of 3 my and an OJPvolume of 24 × 106 to 65 × 106 km3; Tarduno et al.,1991).

However, besides the 90± 4 Ma event, the recognitionof OJP-related lavas with ages of>111–115 Ma coveringFig. 6. Initial ( 208Pb/204Pb)t (a) and ( 207Pb/204Pb)t (b) vs ( 206Pb/204Pb)textensive areas in the Nauru and East Mariana basinsfor central Malaita, southern Malaita, and Santa Isabel basalts. Data

for OJP basement drill sites are indicated by fields with bold outlines. (e.g. Castillo et al., 1994) may suggest a lower overallData sources for the other fields and mantle sources are the same as average emplacement rate (see also Ito & Clift, 1998; Itoin Fig. 5. Th/U values used to adjust the mantle source fields to

& Taira, 2000). Assuming that the 90± 4 Ma event wasestimated 120 Ma positions are 2·25 for Pacific MORB (W. M. White,1993), 3·0 for the Hawaiian source (Tatsumoto, 1978), and 3·2 for the volumetrically important, was characterized by similarMangaia Group (Chauvel et al., 1992). 238U/204Pb values used are as eruption rates to those of the >122 Ma event, and thatin Fig. 5. Analytical errors on ( 208Pb/204Pb)t are about the size of the

magmatism in the period between 94 and 119 Ma wassymbols and those on ( 207Pb/204Pb)t are about twice the height of theminor, the maximum total duration of plateau em-squares.placement in the two eruptive events would be 14 myand the minimum average emplacement rates wouldhave been 1·9 km3/yr and 4·4 km3/yr for off- and on-

Ti, and Y relative to others from the Kwaio Anticline. ridge settings, respectively. These values are similar toHowever, the incompatible element patterns of the Units maximum estimates for the Kerguelen PlateauC–G flows at Site 807, the Site 289 basalt, and most (3·4–5·4 km3/yr; Coffin & Eldholm, 1994) but still muchSanta Isabel flows are similar to those of most central higher than estimates for other large igneous provinces.Malaitan Kwaimbaita Fm. lavas. With the exception of For average emplacement rates to fall as low as valuesevolved sample ML476, Kwaimbaita-type lavas exhibit estimated for the Deccan Traps (0·3–2·4 km3/yr; Coffineither no peak or a variably positive peak at Sr, in & Eldholm, 1994), North Atlantic Tertiary Provincecontrast to the trough at Sr seen for all Malaitan Singgalo (0·6–2·4 km3/yr; Coffin & Eldholm, 1994), or ShatskyFm. lavas. Rise’s South High (1·7 km3/yr; Sager & Han, 1993), for

example, the observed bimodal age distribution on theOJP would have to be an artifact of the few basementlocations currently sampled, such that additional pulses

DISCUSSION or relatively continuous volcanism occurred in theBimodal ages and emplacement rates 86–126 Ma period, possibly for up to 39 my. In the latter

case, the estimated minimum average emplacement ratesThe combined age data now available for the plateauwould fall to 0·7 and 1·6 km3/yr. If the geochemicallyprovide no evidence of a lateral age progression and thusOJP-like volcanism in the East Mariana Basin (0·25 ×an Icelandic-type origin by progressive lateral accretion.

Instead, data for the widely separated drill-hole and island 106 km3) and Nauru Basin (0·85× 106 km3) is included,

17


Fig. 7. Major element oxides vs MgO (wt %) comparing data for central Malaita with those of southern Malaita, Santa Isabel, and the drillsites. Symbols are as in Figs 5 and 6. Sample ML68 is a basalt (Singgalo type, as determined by its isotopic ratios) from northernmost Malaita(Tejada et al., 1996). The four chemical sub-types of lavas from the drill sites, southern Malaita and northeastern Santa Isabel (Mahoney et al.,1993; Tejada et al., 1996) defined previously on the basis of TiO2 contents are indicated in (d) by continuous outlines: A-type, C-G-type, 803-type, and high-Ti Sigana (data for Site 803 and high-Ti Sigana basalts not shown). The dashed outlines in (d) indicate two newly recognizedmagma types: low-Ti Singgalo Fm. lavas (upper field) and low-Ti Kwaimbaita Fm. (lower field).

the estimated total crustal volume associated with the amount of compositionally OJP-like tholeiitic basalt mag-matism younger than >90 Ma. We also note that anOJP for off-ridge and on-ridge emplacement increases

to 28·1 × 106 (Neal et al., 1997) and 62·4 × 106 km3 unknown volume of crust was added to the plateau at>44 Ma, as recorded in the alkalic Maramasike Fm. of(Coffin & Eldholm, 1994), respectively. In this case, the

average emplacement rates would vary from 1·6 to Malaita (Tejada et al., 1996). If the extent of any of theseevents turns out to have been significant, then estimates3·5 km3/yr if magmatism were episodic (total of 6, 4, and

8 my duration in the >122, >111–115, and >90 Ma of overall average emplacement rates would be reducedfurther.events). The >62 and >34 Ma events documented in

San Cristobal (Birkhold-VanDyke et al., 1996; Birkhold Estimates of average emplacement rates will improveas more samples, 40Ar–39Ar dates, and crustal thicknesset al., in preparation) provide evidence of an unknown

18


Fig. 8. Comparison of the primitive-mantle-normalized incompatible element abundances of Singgalo-type lavas with Singgalo Fm. lavas ofthe Kwaio Anticline (a) and Kwara’ae Anticline (b), and of Kwaimbaita-type lavas with Kwaimbaita Fm. lavas of the Kwaio Anticline (c)showing averages and total ranges (shaded envelopes). The average and range for the Kwara’ae Anticline data exclude low-Ti samples KF1 andKF32. Kwaimbaita Fm. average and range exclude ML475 and 476. Average patterns for Singgalo-type lavas of Site 807, southern Malaita,and northeastern Santa Isabel are shown in (a) and (b), whereas average patterns for Kwaimbaita-type lavas from Site 807, Site 803, andnortheastern Santa Isabel are shown in (c). Normalizing values are from McDonough & Sun (1995).

data become available but, so far, all estimates are much on a scale not seen in either modern hotspots or otherlarge igneous provinces.higher than those calculated for modern hotspots

such as Iceland and Hawaii (0·12–0·24 km3/yr and0·03–0·16 km3/yr, respectively; R. S. White, 1993). Pres-

Petrogenetic implications of chemicalent indications are that the contribution of post-90 Mavariationsmagmatism to the plateau’s volume was relatively minorPhase equilibria and possible fractionation paths from(Mahoney et al., 2001). If further sampling confirms thatmultiple parental meltsthe bulk of the OJP was formed at >122 Ma or in two

events at>122 and>90 Ma, it will be difficult to avoid Major element compositions of the majority of OJP lavascoincide approximately with the low-pressure cotectic,the conclusion that the OJP represents melting of mantle

19


Fig. 9. (a) A projection of OJP data from diopside into the planeolivine–anorthite–quartz, assuming Fe3+/Fe2+ = 0·15, together withexperimental data represented by lines with dots at the ends (Herzberg& O’Hara, 1998, and references therein). The broken lines connectingthe dots are the trace of liquids formed by invariant melting ofplagioclase, spinel, and garnet lherzolites at the pressures indicated inGPa. All OJP data project between the 1× 10−4 GPa (1 atm) cotectic[Liq + ol + plag + cpx] of Libourel et al. (1989) and the 1 × 10−4

to 1 GPa initial melt track. The 0·8 GPa cotectic [Liq + ol + plag+ cpx] is from Grove et al. (1992). Continuous lines with half-arrowslabeled a–b demonstrate the effect of olivine fractionation. The dashedline with half-arrowhead labeled c represents inferred olivine-controlledfractionation responsible for some OJP lavas. (b) A projection of theOJP data from olivine into the plane CaSiO3–MgSiO3–Al2O3 (Herzberg& O’Hara, 1998), which contains pyroxenes and garnet. Lines labeled0·2, 0·8, and 1·0 are cotectics [Liq+ ol+ plag+ cpx] at the pressuresindicated in GPa (Grove et al., 1992).Β, invariant points at 1× 10−4,1, and 2 GPa from Longhi (1987) and Walter & Presnall (1994).

liquid + olivine + plagioclase + clinopyroxene, con-sistent with the mineral phases observed and the ophiticto subophitic texture of clinopyroxene and plagioclase inthe phyric lavas. The differentiation trajectories (arrowsa–c) in Fig. 9a illustrate olivine-controlled fractionationtrends. The spread of data between the 1 × 10−4 GPa(1 atm) and 0·8 GPa (8 kbar) cotectics (Fig. 9a and b)demonstrates the importance of crustal-level fractionationprocesses.

Tab

le5

:A

vera

gees

tim

ated

min

imum

empl

acem

ent

rate

sfo

rth

eO

nton

gJa

vaP

late

au

Volu

me

(km

3 )M

axim

um

bA

ge

(Ma)

Max

imu

mE

mp

lace

men

tra

tes

(km

3 /yr)

Ass

um

pti

on

Min

imu

ma

(on

-rid

ge)

du

rati

on

(my)

(off

-rid

ge)

Min

imu

mM

axim

um

(off

-rid

ge)

(on

-rid

ge)

27·0×

106

61·3×

106

119–

125

64·

510

·2O

ne

volc

anic

even

tfo

rth

ew

ho

lep

late

au

27·0×

106

61·3×

106

86–1

2539

0·7

1·6

Co

nti

nu

ou

ser

up

tio

nfr

om

122

Ma

even

tto

90M

a

even

t

27·0×

106

61·3×

106

119–

125

141·

94·

4Tw

oep

iso

dic

even

tsat

122

Ma

and

90M

a

+86

–94

28·1×

106

62·4×

106

119–

125

181·

63·

5E

pis

od

ic12

2M

aan

d90

Ma

even

ts,

plu

s

111–

115

Ma

Nau

ruB

asin

and

Eas

tM

aria

na

Bas

inev

ent

+11

1–11

5

+86

–94

Volu

me

esti

mat

esar

efr

om

a Nea

let

al.

(199

7)an

dbC

offi

n&

Eld

ho

lm(1

994)

.

OJP lavas also may be inferred to be derivative liquidsthat have undergone significant olivine fractionation at

20


greater pressures; yet we do not see direct evidence for in the Kwaimbaita-type lavas at a given mg-numbersuggest larger fractions of partial melting than for thesuch fractionation because the effects of low-pressure

fractionation effectively dominate the major element sig- Singgalo-type basalts. Previous major and trace elementmodeling indicates that both groups of rocks were formednature of the lavas (see Neal et al., 1997). Although some

Kwaimbaita lavas may have differentiated from>1 GPa by large total amounts of partial melting (most estimatesare in the 18–30% range, assuming peridotite sources),melts (arrow a in Fig. 9a), the majority of the data appear

to fall at the ends of olivine-control trajectories originating but suggest that the Kwaimbaita basalts represent severalpercent more melting (Tejada et al., 1996; Neal et al.,from melts formed between 1 and 4 GPa (arrow b).

Interestingly, trace element modeling also requires 3·5– 1997; Tejada, 1998).Alternatively, can the difference in chemical com-4·0 GPa initial melting pressures to fit the OJP data

(Tejada, 1998; see also Mahoney et al., 1993; Neal et al., positions between the Kwaimbaita and Singgalo magma1997). In any case, the diagram indicates that, if they types be a function of source differences? The values forformed from peridotite source mantle, the erupted OJP many incompatible element ratios among rocks fromlava compositions are significantly different from primary different locations on the OJP overlap, particularly whenmagma compositions (e.g. Herzberg & O’Hara, 1998). interlaboratory variation and variable analytical precision

Some data for Singgalo-type lavas fall at the end of are considered. In diagrams involving ratios of elementsthe 4–6 GPa olivine-control trend (dashed arrow c is the with significantly different bulk partition coefficients and6 GPa olivine control), similar to most Hawaiian lavas having higher analytical precision, such as Ce/Sm, Zr/(Herzberg & O’Hara, 1998). Although this type of dia- Nb, Gd/Yb, and Zr/Y (Fig. 10b, c, e and f ), a weakgram, derived from isobaric experimental data, provides correlation with La abundance is observed. On the otheronly a rough idea of equilibration pressures of polybaric hand, ratios of elements with similar incompatibilities,natural melts, the similarity with Hawaiian tholeiites is such as La/Nb and Nb/Th (Fig. 10a and d), show noperhaps not too surprising considering that the Singgalo clear covariation with La (or with isotopic ratios). ThisFm. was erupted at the end of the 122 Ma episode, when observation is consistent with both the generally similarthe lithosphere of the OJP had probably been thickened shapes of primitive-mantle-normalized incompatible ele-significantly by the emplacement of the earlier Kwaim- ment patterns for Singgalo- and Kwaimbaita-type lavasbaita-type magmas, cumulates, and melt-depleted res- (Fig. 8) and the generally slightly higher abundances ofidues. A deeper origin for Singgalo-type magmas is also incompatible elements in the Singgalo-type lavas, andconsistent with trace element modeling suggesting higher probably reflects slightly lower degrees of melting for theinitial pressures of melting for Singgalo-type magmas latter. Because ratios of incompatible trace elements with(Tejada, 1998). similar incompatibilities, such as Nb/Th, and La/Nb,

Another important indication of polybaric fractionation are not changed significantly by moderate amounts ofis given by the Al:Ti ratios measured in clinopyroxene, magmatic differentiation or by variations in melt fractionpresent both as a phenocryst and groundmass phase, in at large total amounts of partial melting, they must mainlylavas from central and northern Malaita. Values range reflect the extent of variation in mantle source regions.between 4 and 10, implying pressures of formation from Figure 10a and d suggests that although the sources of>1 × 10−4 GPa (>1 atm) to a maximum of 1 GPa OJP lavas had distinct isotopic compositions, they had(Babbs, 1997). This finding is consistent with the position very similar ratios of highly incompatible trace elementsin Fig. 9 of the bulk of the whole-rock major element (see also Neal et al., 1997).data, which fall between the 1 × 10−4 GPa and 0·8GPa cotectics.

Bimodal isotopic compositions: a vertically zoned mantlesource for the OJP?

Source variation vs melting processes The combined isotopic data now available indicate thata mantle source containing material of two distinct com-Among the major elements (Fig. 7), TiO2, P2O5 (not

shown) and, to a lesser extent, Fe2O3∗ show a clear positions, the Singgalo and Kwaimbaita types, dominatedmagma production, at least for the later stages of vol-distinction between Kwaimbaita-type and Singgalo-type

lavas at similar mg-number. This difference could po- canism represented by basalts in the upper few kilometersof OJP crust. Importantly, there is no evidence of eithertentially reflect the composition of the mantle sources or

be a result of differences in the amounts and/or pressures magma or source mixing between the two types duringthe>122 Ma event. It should be noted that the bimodalof partial melting. TiO2, for example, is sensitive to the

amount of partial melting (e.g. Hirose & Kushiro, 1993), distribution of data on the Pb–Pb isotope diagrams (Fig.6) was less sharply defined in previous studies becausebehaving as an incompatible element during mantle

melting and attaining the highest concentrations in small- most Pb isotope ratios were not age-corrected (the age-corrected Pb isotopic data in Table 4 are the first fordegree partial melts. The lower concentrations of TiO2

21


Fig. 10. Selected incompatible element ratios vs La and �Nd(t). The �Nd(t) value of 5·5 approximates the boundary between the two chemicaltypes (also see Fig. 4). Fine dashed lines separate data for the Singgalo and Kwaimbaita formations and error bars represent typical analyticaluncertainties for data of this study.

which parent and daughter nuclides were both measured Mauna Loa lavas have lower 207Pb/204Pb and, par-ticularly, 208Pb/204Pb.on the same sample splits by isotope dilution).

A two-component mantle source also has been inferred, The geochemical stratigraphy in central Malaita andat Site 807 indicates a temporal change in the compositionfor example, for the Hawaiian hotspot, currently the

world’s most active, with end-members represented by of the OJP mantle source. At Mauna Loa volcano, asomewhat similar shift from earlier (i.e. in the prehistoricshield lavas of Kilauea and Koolau volcanoes (e.g. Rhodes

& Hart, 1995; Chen et al., 1996). Intriguingly, the Kwaim- lavas) Kilauea-like to more ‘typical’, lower-206Pb/204Pband lower-�Nd Mauna Loa compositions in the historicalbaita and Kilauea isotopic ranges overlap significantly

(Figs 5 and 6). Koolau values range to lower �Nd and lavas (Kurz et al., 1995) coincided with declining magmaproduction (Moore et al., 1990). The analogy for the OJP206Pb/204Pb than yet found for the OJP but Singgalo-type

isotopic values are close to, or overlap with, those for is limited, however, because the different Hawaiian shieldvolcanoes and/or lava sequences within a single shieldhistorical lavas of Mauna Loa volcano, except that the

22


appear to represent different amounts of mixing between the starting-plume or plume-head model (e.g. Richardset al., 1989; Griffiths & Campbell, 1990). Some laboratorythe two end-member components, whereas no mixingand numerical experiments suggest that, for all but thebetween Kwaimbaita- and Singgalo-type sources orlargest plume heads, significant amounts of ambient,magmas is observed in the >122 Ma lavas. Rather, themostly lower mantle may be entrained during the ascentshift from Kwaimbaita-type to Singgalo-type com-of a plume head from a thermal boundary layer withinpositions is abrupt in the two locations where it hasthe mantle, possibly from the core–mantle boundary, andbeen found (i.e. Site 807 and Malaita). The lack oftherefore plume heads are predicted to be compositionallyintermixtures between the two OJP isotopic groups, theirzoned or layered (e.g. Griffiths & Campbell, 1990). Thus,consistent stratigraphic position in lavas emplaced atin open-ocean areas such as that in which the OJPessentially the same time in two locations >1600 kmformed (e.g. Neal et al., 1997), where complications arisingapart, and the probably small amount of drift of thefrom a plume’s interaction with continental lithosphereplateau during the 125–100 Ma period (Neal et al., 1997)are absent, the isotopic compositions of the volcanicsuggest that the two mantle reservoirs in the OJP sourceproducts are predicted to reflect the plume-source andwere much larger in size than the scale of melting and,some amount of entrained lower mantle. The differentin the melting region, may have been vertically separatedocean-island-like isotopic compositions of the OJP lavasfrom one another.are not inconsistent with these predictions, as the Sing-galo- and Kwaimbaita-type isotopic groups would rep-resent internally homogeneous (relative to the scale of

Modeling the OJP’s mantle source melting) mantle regions composed of slightly differentIn recent years, most attempts to explain large igneous proportions of plume-source mantle and entrained, pre-provinces have invoked mantle plumes. An exception is sumably lower, mantle.the plate reorganization model (e.g. Anderson et al., 1992; Although interpretation is very model-dependent, theSmith & Lewis, 1999), in which flood basalt volcanism Kwaimbaita-type basalts in the framework of this classis initiated by plate reorganization above anomalously of models would more probably reflect the greaterhot, shallow, non-plume mantle. Initially, melting is as- amount of plume-source-derived mantle. First, both insumed to tap a hypothesized widespread, shallow mantle central Malaita and at Site 807, the thickness of thelayer with ocean-island-like isotopic and trace element earlier, Kwaimbaita-type lavas is much greater than thatcharacteristics [termed the ‘perisphere’ by Anderson of the later, Singgalo-type lavas, which appears consistent(1996)]; this layer is later replaced by upwelling mid- with laboratory and numerical models that suggest largeocean ridge basalt (MORB)-type mantle that lies beneath plume heads will contain much more plume-source ma-it. The model thus predicts that MORB-like isotopic terial than entrained mantle (e.g. Griffiths & Campbell,compositions should dominate the signature of late-stage 1990). Second, the chemical compositions of Kwaim-lavas on oceanic plateaux. Although the OJP’s>122 Ma baita-type basalts indicate derivation by somewhat largerevent, in particular, occurred relatively close in time to amounts of partial melting of peridotite, consistent witha major reorganization of the Pacific plate (Neal et al., hotter, plume-source-dominated mantle (see above).1997), the complete absence, thus far, of isotopically Third, the >90 Ma eruptive event, which may haveMORB-like lavas in the >122 Ma (and >90 Ma) OJP been centered on the plateau’s eastern lobe or salientcrustal section does not support the plate reorganization (Tejada et al., 1996; Neal et al., 1997), produced lavasmodel. Rather, the Singgalo Fm. lavas are even less with Kwaimbaita-like compositions at Site 803, SantaMORB-like than the earlier Kwaimbaita Fm. basalts. Isabel, and San Cristobal, whereas >90 Ma Singgalo-

Another class of model posits that oceanic plateaux type lavas have been found only in San Cristobal. Fourth,are formed gradually over tens of millions of years by the 111–115 Ma Nauru, East Mariana, and Centralmore or less steady-state, ridge-centered or near-ridge Pacific Basin basalts, which appear to be closely relatedplumes, rather like the comparatively small Icelandic to emplacement of the OJP, have isotopic ratios thatPlateau or Galapagos Platform and its associated aseismic overlap with those of the Kwaimbaita Fm. (e.g. Figs 5ridges (e.g. Mahoney & Spencer, 1991; Ito & Clift, 1998). and 6). Thus, the Kwaimbaita-type signature appears toHere, involvement of MORB-type mantle also might be be expressed much more commonly than the Singgaloexpected, particularly in locations distant from the plume type, suggesting that the former was volumetrically morecenter. Perhaps more importantly, this model finds no abundant in the OJP mantle source.support in the currently available age data for OJP However, several critical aspects of the OJP do notbasement lavas, which indicate that eruption occurred obviously conform to the starting-plume model (see alsoin short-lived pulses >30 my apart. Mahoney et al., 1993; Tejada et al., 1996; Neal et al., 1997;

By far the most frequently applied plume-based model Sheth, 1999). The model predicts rapid emplacement ina single event; for the OJP this would require that thefor large igneous provinces in the last decade has been

23


>90 Ma event was volumetrically negligible and caused plume peridotite assumed to surround pockets of recycledcrust must have Pb isotopic values as low as or lowerby magmatism related to the plume tail that followed the

plume head. At present, the magnitude of the >90 Ma than observed in the OJP basalts. Figure 11a–d illustratesthat it is indeed possible to approximate the OJP’s isotopicepisode remains unknown, but it was clearly widespread,

showing up in locations as far apart as Site 803, Site characteristics by mixing 0·5–20% partial melts of peri-dotite, assumed here for simplicity to have model prim-288, San Cristobal, and Santa Isabel. Also, plume-tail-

related volcanism following the plume-head phase is itive-mantle isotopic and elemental compositions, with a100% melt of recycled upper oceanic crust (eclogite), thepredicted to form a post-plateau chain of seamounts

adjacent to a plateau as plate motion moves the plateau isotopic characteristics of which have been modifiedby seafloor aging and subduction-related dehydrationaway from the hotspot, but the seafloor adjacent to the

OJP lacks a post-90-Ma chain; if one ever existed, it processes (see Table 6 for the model compositions usedin constructing Fig. 11). The calculations show that anmust have been obliterated by subduction along the old

Solomon trench. Recent numerical modeling suggests 80–98% proportion of eclogite melt is required to broadlyreproduce the 206Pb/204Pb, Nd, and Sr isotopic values ofthat a different type of large plume head may form at

mid-mantle depths beneath the 660 km discontinuity, OJP lavas if the plume-source peridotite melts only to avery small extent (0·5%), consistent with Cordery et al.’srather than above it or the core–mantle boundary (Cse-

repes & Yuen, 2000). Such plume heads may not be (1997) ‘best’ melt volume scenario. However, the amountof peridotite partial melting that accompanies completeaccompanied by very long-lived tails (D. Yuen, personal

communication, 1999); if so, this model may account for melting of the eclogite generally might be expected tobe somewhat higher, considering that the eclogite itselfthe lack of a post-OJP seamount chain. However, this

and other plume-head models also predict that substantial ultimately represents MORB-type crust formed by atleast several percent of partial melting of (non-plume)uplift accompanies plateau emplacement, with large areas

at shallow depths or even emergent; yet paleodepths peridotite. If the plume-source peridotite melts to agreater extent (e.g. 20% in the example illustrated in Fig.appear never to have been shallow at any of the locations

sampled thus far in Malaita, Santa Isabel, or at any of 11), a smaller proportion (only 40–60% in the example)of recycled oceanic upper crust can produce the OJPthe drill sites (e.g. Neal et al., 1997; Ito & Clift, 1998;

Michael, 1999; Ito & Taira, 2000; Mahoney et al., 2001). isotopic compositions. In all cases, with the primitive-mantle and eclogite Pb isotope values used for the figure,Furthermore, if the>90 Ma event was relatively minor,

a truly enormous plume head is required to fuel melting the 207Pb/204Pb ratios of the postulated hybrid magmasare much lower than observed for the OJP lavas for anyof peridotite at a rate and volume capable of building

the huge OJP in only a few million years or less at given 206Pb/204Pb value. Of course, the plume peridotiteneed not have primitive-mantle isotopic (or elemental)>122 Ma (e.g. Coffin and Eldholm, 1994).

More recently, several studies have modified the start- values; a more EM-1-like end-member, for example,provides higher relative 207Pb/204Pb although still yieldinging-plume model by proposing that plume heads, al-

though consisting principally of peridotite, may contain reasonable fits in Fig. 11a–d. Even if the plume peridoditeis primitive mantle, this may not present much of asignificant amounts of embedded eclogitic material de-

rived from subducted oceanic crust (e.g. Cordery et al., problem because of the large uncertainties in, and thelarge range of, estimates of primitive-mantle Pb isotopic1997; Yasuda et al., 1997). The idea behind this ‘composite

diapir’ model is that the eclogite melts extensively before ratios (dashed field in Fig. 11e; e.g. Galer & Goldstein,1996).the peridotite begins to melt, but cannot escape the

source until peridotite melting allows large-scale melt Model incompatible element patterns generated bymixing batch melts of primitive mantle and recycledconnectivity, at which point large volumes of melt may

be released cataclysmically. Yasuda et al. (1997) showed oceanic crust are shown in Fig. 12. Patterns for unalteredand altered MORB modified by dehydration in a sub-that the major element compositions of hybrid magmas

derived by melting of such mixed sources could be duction zone were estimated using the trace elementmobilities derived by Kogiso et al. (1997). The calculatedcomparable with those of continental flood basalts, im-

plying that some flood basalts could be near-primary mixtures do not reproduce the patterns of OJP lavas toosuccessfully. Models assuming a purely peridotitic sourcemagmas.

Intra-oceanic plateaux such as the OJP should be the give a better fit to the data (see Tejada et al., 1996; Nealet al., 1997; Tejada, 1998). In particular, the patterns ofbest places to test the composite diapir model using

isotopes, trace elements, and major elements, because OJP lavas lack the U and Rb enrichment expected foraltered, recycled MORB and have different overall shapesdirect continental lithospheric influence on magma com-

positions is lacking. Because subducted, recycled oceanic than the model patterns. Slightly better fits are achievedif the recycled crust is assumed to be unaltered whenupper crust is expected to develop high 206Pb/204Pb ratios

on time scales of >108 yr (e.g. Kogiso et al., 1997), the subducted and dehydrated. Although none are very good,

24

TE

JAD

Aet

al.O

RIG

INA

ND

EV

OL

UT

ION

OF

ON

TO

NG

JAV

APL

AT

EA

U

Table 6: Isotope systematics for one model of recycled oceanic crust (eclogite) and model primitive mantle in the OJP plume

source

206Pb/204Pb 207Pb/204Pb 208Pb/204Pb 87Sr/86Sr 143Nd/144Nd �Nd

MORB values at t = 0 18·637 15·505 38·126 0·70266 0·51311 +9·1

MORB source at t = 300 my P/D 5 0·04 12·5 0·02 0·24

Initial ratios 18·40 15·49 37·94 0·70257 0·51264 +7·5

MORB crust after 100 my P/D 11 0·08 29·9 0·018 0·215

(unaltered) Ratios 18·57 15·5 38·09 0·7026 0·51278 +7·7

Recycled MORB at 120 my P/D 103 0·75 123·6 0·195 0·264

(altered) Ratios 19·86 15·56 38·58 0·70282 0·51291 +8·4

PM at t = 0 Ratios 17·781 15·42 38·149 0·705 0·51264 0

PM at t = 120 my P/D 8·2 0·06 32·8 0·085 0·1967

Ratios 17·63 15·41 37·95 0·70486 0·51249 0

Conc. 0·071 0·071 0·0795 19·9 1·25

MORB, mid-ocean ridge basalt; PM, primitive mantle, P/D, parent–daughter ratios: 238U/204Pb for 206Pb/204Pb, 235U/204Pb for 207Pb/204Pb, 232Th/204Pb for 208Pb/204Pb,87Rb/86Sr for 87Sr/86Sr, and 147Sm/144Nd for 143Nd/144Nd. �Nd = 0 corresponds to 0·512254 at 300 my, 0·512383 at 200 my, and 0·512486 at 120 my. Parent–daughtervalues are from W. M. White (1993) and Mahoney et al. (1994) for southern East Pacific Rise MORB, and from Jacobsen Wasserburg (1980), McCulloch Black(1984) and Galer Goldstein (1996) for primitive mantle. Concentration values and parent–daughter ratios for recycled MORB are from Kogiso et al. (1997) andreferences therein.

25


Fig. 11. Example of isotopic mixing curves for two possible combinations of plume peridotite and recycled oceanic upper crust (eclogite) end-members in the OJP mantle source. The plume peridotite is assumed to have primitive mantle compositions and the recycled oceanic crust isassumed to have formed as MORB at 300 Ma from a Pacific MORB-type source, then ‘aged’ on the seafloor for another 100 my before beingsubducted, and then to have evolved in the mantle for 80 my before being incorporated into the OJP plume diapir (see Table 6). Theconcentration and parent–daughter values used are for the less-altered basaltic MORB crust of Kogiso et al. (1997). Two extreme cases areillustrated: one case in which the recycled crust melts completely and mixes variably with a 0·5% partial melt from primitive mantle (PM), andthe other in which a 100% melt of recycled crust is mixed with a 20% partial melt of primitive mantle in various proportions. The first caseindicates that OJP isotopic compositions can be approximated by 80–98% proportions of recycled oceanic crust melt, mixed with very smallpartial melts of primitive mantle. The second case indicates that if peridotite in the OJP plume mantle is capable of partially melting up to 20%,only 40–60% of recycled oceanic crust is required to explain OJP-type 206Pb/204Pb, Nd, and Sr isotopic values in the hybrid magmas. Comparablefits also can be achieved by assuming an older age for the recycled MORB end-member. However, all models fail to explain the OJP data fieldsin the207Pb/204Pb vs 206Pb/204Pb diagram, a feature that may be alleviated by choosing a different (yet allowable) primitive-mantle Pb isotopecomposition [shown as dashed fields in (d) and (e)] or a non-primitive mantle composition. S. EPR MORB, southern East Pacific Rise MORB;OJP-S, OJP Singgalo; OJP-K, OJP Kwaimbaita. Data sources are as in Table 6.

the best fits are obtained with>100% melting of recycled from melting experiments on fertile peridotite KLB-1(columns 2, 5, and 8 in Table 7; data of Hirose &oceanic crust and high proportions of crustal melt in

the hybrid magma (>40% or 90–100%, respectively, Kushiro, 1993). Table 7 lists the compositions of hybridmagmas comprising large-degree partial melts (>20%)assuming the plume peridotite partially melts to 20% or

0·5%), as in the isotopic modeling. These proportions of fertile peridotite and total melts of recycled upperoceanic crust [taken to be a 100% melt of crust plus 1%are similar to those hypothesized for continental flood

basalts by Yasuda et al. (1997) from major element peridotite melt, following Yasuda et al. (1997); their otherhypothetical mixes have higher SiO2 and Al2O3 contentsconsiderations.

To estimate the major element compositions of hybrid than most OJP lavas]. Proportions of crustal melt in themixtures shown are 0%, 40%, and 60%, and results forprimary OJP magmas, we used Yasuda et al.’s (1997)

data on melted MORB (column 1 in Table 7) and results melts in the depth range 2·5, 2·0, and 1·5 GPa are

26


Fig. 12. Calculated primitive-mantle-normalized incompatible element compositions for batch melts of a composite-diapir source containingprimitive mantle (PM) and recycled, aged (altered) and dehydrated upper oceanic crust (MORB). Mixing proportions are based on isotopicconsiderations (see Fig. 11). The patterns with no markers represent the average compositions of Kwaimbaita Fm. (bold filled) and Singgalo Fm.(light dashed) basalts, and average unaltered, but subducted and dehydrated, normal MORB (bold dashed pattern). The top and bottom panelsshow the trace element patterns produced if a 100% melt of recycled, altered crust is mixed with 0·5% and 20% batch partial melts, respectively,of plume peridotite in various proportions (X values, with X = 1 corresponding to no plume peridotite input). The bold filled gray patternswith no markers represent the mixtures involving unaltered, dehydrated MORB that best fit the patterns of OJP lavas [they correspond to X=0·98 (top panel) and X = 0·40 (bottom panel)]. Data sources are Mahoney et al. (1998) for altered MORB (average of data for DSDP Sites 235and 245) and Sun & McDonough (1989) for average normal MORB and primitive mantle compositions. The altered and average normalMORB compositions were adjusted to dehydrated compositions using the mobility factors determined by Kogiso et al. (1997). Distributioncoefficients used are from Arth (1976), Green et al. (1989), McKenzie & O’Nions (1991), Hart & Dunn (1993), Kennedy et al. (1993), Bedard(1994), the compilation of Green (1994), Hauri et al. (1994), Gurenko & Chaussidon (1995), and Xie et al. (1995).

presented. The mixtures show a range of SiO2, Al2O3, followed by up to 15% fractionation of the olivine +clinopyroxene + feldspar.] The model fractionationTiO2, and CaO contents, which overlap with those of

OJP lavas, except that hybrid primary magmas containing paths, although somewhat lower in Fe2O3∗ and CaOand higher in TiO2 than the majority of OJP lavas,large proportions of peridotite melt have lower Fe2O3∗

and much higher MgO contents (11–13 wt %) than seen intersect the OJP range. If such fractionation is consideredin the trace element modeling in Fig. 12, the principalfor any OJP basalt. The difference in MgO contents can

be removed by fractionating the model primary magmas effect is to increase the abundances of the incompatibleelements. For example, 30% fractionation of an as-to levels seen in OJP basalts. Fractionation modeling

using the ‘MELTS’ program (Ghiorso & Sack, 1995) is semblage consisting of 15% clinopyroxene, 75% plagio-clase, and 10% olivine leads to a factor of 1·4 and 1·3generally consistent with up to 10% fractionation of

olivine and spinel, followed by 10–40% removal of clino- increase, respectively, in the abundance of elements withbulk distribution coefficients of 0·04 (e.g. Th, Dy) andpyroxene and plagioclase, at pressures varying between

>0·05 and 0·2 GPa (Fig. 13). [An exception is the path 0·20 (e.g. La). This increase slightly improves the con-cordance of the best-fit model patterns with those of thecalculated for the pure crustal melt of Yasuda et al. (1997),

which entails up to 15% olivine+ feldspar fractionation OJP averages, particularly from Th to Dy. In view of

27


the uncertainties in source composition, fractionation than the interior. However, the mantle near 660 kmcrossed by the plume would be expected to contain muchhistory, modification of some elemental abundances by

post-eruptive alteration, and the ability of the MELTS less eclogite at>90 Ma than at>122 Ma, because muchof it would have already been dragged upward by theprogram to reproduce real-world processes, it appears

that melting in a composite diapir may give rise to plume head shortly before >122 Ma. Finally, if theproportion of eclogite involved in producing OJP magmasbroadly OJP-type magmas under some conditions.

In summary, we cannot rule out the composite diapir was high, as the model suggests, then the plateau’s hugesize requires a very large amount of recycled oceanicmodel for the OJP on the basis of isotopic or chemicalcrust to have been concentrated in a relatively smalldata alone, given the uncertainties in end-members andmantle region, so as to become incorporated in a risingin effects of alteration and subduction on oceanic upperplume head. If we assume, for example, that 50 × 106crust. A major advantage of this model, relative to thekm3 of eclogite melted, this value translates to an eclogite‘simple’ plume-head model, is that it does not requirelayer of 100 km thickness and 399 km in radius; assumingan anomalously hot and huge plume head to producefurther that such a layer was derived ultimately from thethe rate and volume of melting required for continentalbasaltic upper>2·5 km of the oceanic crust, an immenseand oceanic flood basalts. However, in the compositeoriginal area of subducted seafloor (4472 km× 4472 km)diapir model temperature variations and variations inis required. In short, it appears that, as with the ‘simple’the distribution of eclogite within the plume head stronglyplume-head model, a number of contortions are requiredaffect the rate and volume of melting, and thus thefor the composite-diapir model to work for the OJP.geochemistry of magmas (Cordery et al., 1997). Yet one

of the most outstanding characteristics of the OJP basalts,thus far, is their small overall range of geochemicalvariation. Furthermore, the stratigraphic succession of

OJP’s crustal growth and mantle sourceSinggalo-type lavas above Kwaimbaita-type lavas is thestructure: a synthesissame in both locations (>1600 km apart) where bothCrustal growth and structurelava types have been sampled in a single section. The

calculations of Cordery et al. (1997) suggest that only the Schematic diagrams of possible OJP crustal and mantlematerial in the uppermost portion of the plume head source structure during the>122 Ma event, based on the

results of this study and previous geochemical modelingwill melt significantly. Following Griffiths & Campbell(1990), they assumed the plume head to have a hot (Mahoney et al., 1993; Tejada et al., 1996; Neal et al.,

1997; Tejada, 1998), are shown in Fig. 14. An initialcylindrical core (axis) with a thin cap of hot plume-source-dominated material along its upper boundary; the emplacement somewhat off-axis (Mahoney & Spencer,

1991; Coffin & Gahagan, 1995; Gladczenko et al., 1997),remainder is composed of a cooler mixture of plume-source and entrained lower-mantle material. If so, the and a peridotitic (Fig. 14a) or composite (Fig. 14b) plume

head are assumed, despite the first-order discrepanciesKwaimbaita–Singgalo succession would appear to re-quire an abundance of eclogite within the upper boundary noted in ‘Modeling the OJP’s mantle source’ between

observed features of the OJP and predictions of all currentlayer, which would melt first to produce the Kwaimbaita-type lavas. Subsequent melting of plume mantle con- plume-head models. Because of these limitations, the

figure should be taken as only a rough working model fortaining a lesser amount of eclogite would produce theSinggalo-type lavas. future studies. It emphasizes the importance of intrusion,

underplating, and cumulate formation (e.g. Neal et al.,One important feature that is difficult to explain witha non-uniform distribution of eclogite within the plume 1997; Ito & Clift, 1998; Ito & Taira, 2000). Modeling

assuming purely peridotite melting suggests that melthead is the observed abruptness of the change from oneisotopically distinct, relatively uniform magma type to generation was polybaric and probably at high potential

temperature and initial pressures of at least 4 GPaanother. Also, most of the >90 Ma OJP basalts studiedthus far are isotopically equivalent to the >122 Ma (Tejada, 1998; and ‘Petrogenic implications of chemical

variations’); under these conditions, picritic, high-MgOKwaimbaita, not Singgalo, lavas, which would requirethat the plume tail, >30 my after the initial plateau- primary melts were probably formed (e.g. Herzberg &

Zhang, 1996; Herzberg & O’Hara, 1998). Exposed sec-forming event, must again contain a high proportion ofeclogite, a proportion very similar to that in the upper tions of the much smaller Caribbean Plateau indeed

document the existence of picritic lavas at deep crustalpart of the starting-plume head at >122 Ma. Yasuda etal. (1997) argued that the eclogite is entrained as the levels (e.g. Kerr et al., 1997). The fact that most OJP

lavas are relatively Fe-rich basalts, differentiated torising plume head crosses an eclogite-rich zone near660 km depth (where eclogite should be neutrally buoy- roughly similar extents, suggests that primary high-MgO

magmas were ponded below the crust or at the Mohoant), which might explain how the upper part of a plumehead could contain more recycled oceanic crustal material (e.g. Cox, 1980; Neal et al., 1997; see Fig. 9). Fractionation

28

TE

JAD

Aet

al.O

RIG

INA

ND

EV

OL

UT

ION

OF

ON

TO

NG

JAV

APL

AT

EA

U

Table 7: Major element compositions of hypothetical hybrid magmas in a composite diapir source

100% crust + KLB-1 2·5 GPa melt KLB-1 2·0 GPa melt KLB-1 1·5 GPa melt OJP lavas

1% peridotite melt F = 0·20 F = 0·22 F = 0·19

X: 1·0 0 0·4 0·6 0 0·4 0·6 0 0·4 0·6

Melt index

(for Fig. 13): 1% P ‘25·4’ ‘25·6’ ‘20·4’ ‘20·6’ ‘15·4’ ‘15·6’

SiO2 49·94 47·92 48·73 49·13 48·26 48·94 49·27 48·72 49·21 49·45 48·60–50·50

TiO2 1·44 0·68 0·99 1·14 0·51 0·88 1·07 0·59 0·94 1·11 0·90–1·80

Al2O3 15·19 13·67 14·28 14·58 13·03 13·89 14·32 15·05 15·11 15·13 13·50–15·90

Fe2O3∗ 10·61 9·32 9·84 10·10 9·68 10·05 10·24 8·31 9·23 9·69 11·90–15·00

MgO 8·22 15·73 12·72 11·22 15·54 12·61 11·15 13·00 11·09 10·13 8·90–6·30

CaO 11·36 10·83 11·04 11·15 10·95 11·11 11·20 12·18 11·85 11·68 10·40–13·20

Na2O 2·55 1·45 1·89 2·11 1·36 1·84 2·07 1·57 1·96 2·16 1·48–3·08

K2O 0·19 0·15 0·16 0·17 0·13 0·15 0·16 0·08 0·12 0·14 0·04–0·26

Total 99·50 99·75 99·65 99·60 99·46 99·48 99·48 99·50 99·50 99·50

The columns in bold type represent end-member compositions. F, fraction of plume peridotite melting; X, fraction of 100% crust + 1% peridotite melt mixturefrom Yasuda et al. (1997) in column 1. Experimental melt compositions are from Hirose & Kushiro (1993). Mixing proportions are based on isotope and traceelement modeling results. FeO contents of the experimental and model melts are converted to Fe2O3∗ for comparison with OJP data. Melt index values are labelsused in Fig. 13.

29


Fig. 13. Major element modeling results for crystal fractionation of mixed crustal eclogite (ex-crust) and peridotite (plume) melts using Ghiorso& Sack’s (1995) ‘MELTS’ program. Parental melt compositions are from Table 7 [e.g. 1% P–0·2 GPa means the melt composition of the 100%crustal + 1% peridotite melt mixture of Yasuda et al. (1997) was fractionated at 0·2 GPa]. The markers indicate 10% fractionation intervals. Itshould be noted that all model runs of the different parental melt compositions at similar pressures of crystallization tend to follow nearly thesame liquid lines of descent (LLD) after >20% fractionation, and thus, for clarity, only a few, representative runs encompassing the full rangeof results at 0·05 GPa, 0·1 GPa, and 0·2 GPa are shown. The modeling shows that, in a general way, the OJP data may be explained byfractionation at variable depths, consistent with other petrogenetic indicators, such as the phase diagram (Fig. 9), discussed in ‘Petrogeneticimplications of chemical variations’ and suggesting the existence of magma chambers at crustal depths.

of olivine and spinel would decrease the density of the shallow-level assimilation of hydrothermally altered crust(Michael, 1999).melts and enable them to ascend and fractionate further

in crustal magma chambers; some magmas undoubtedly Because the portion of the Pacific plate containing theOJP appears to have been near the 125–100 Ma Pacificfroze en route in intrusions, whereas some reached the

surface, leaving behind more gabbroic cumulates (e.g. stage poles and thus drifted little relative to the putativeplume (e.g. Neal et al., 1997), Fig. 14 shows vertical toCox, 1980, 1993). The ‘hidden cumulates’ in the deeper

levels of OJP crust probably represent a significant portion slightly offlapping stacking of >122 Ma volcanic prod-ucts. By the time of the Singgalo Fm. eruptions, the earlierof the high-seismic-velocity lower-crustal layer of the

plateau (Farnetani et al., 1996; Neal et al., 1997). In a volcanic pile and associated cumulates and intrusions hadthickened the OJP’s crust enormously. This change iscomposite diapir scenario, gabbroic cumulates would be

less abundant because the hybrid primary magmas would reflected in the chemical compositions of Singgalo-typelavas, which appear to represent lower mean fractionsbe less Mg rich. The megacrysts and xenoliths in the

‘orbicular’ basalts in central Malaita described in the of peridotite melting and higher pressures of melt se-gregation than the Kwaimbaita basalts (Tejada et al.,section on field characteristics exemplify relatively shal-

low-level gabbroic cumulates. Also, the high Cl/K ratios 1996; Neal et al., 1997; Tejada, 1998). In general, lessertotal amounts of melting would mean decreased magmaof basaltic glasses at Site 807 are suggestive of relatively

shallow magma residence because they indicate some supply to magma chambers, possibly leading to slightly

30


Fig. 14. (a) Schematic crustal and mantle source structure for the OJP, assuming a plume-head origin. Stage 1 (bottom) represents initial>122 Ma eruption of Kwaimbaita-type lavas fed by melting of the upper portion of the mantle source region, shown as a plume head consistingmainly of plume-source-derived peridotite (e.g. Griffiths & Campbell, 1990). Stage 2 (top) illustrates the subsequent emplacement of Singgalo-type lavas atop Kwaimbaita-type lavas at >122 Ma. Much of the plume-head mantle has been melted and the melt erupted, intruded into, orunderplated to the crust, and the crust has been thickened greatly. Magma chambers are situated at crustal levels and/or near the crust–mantleboundary. The Pacific plate is moving relatively slowly with respect to the mantle source region (Neal et al., 1997). (See text for additionalexplanation.) (b) Schematic representation of a composite-diapir mantle source for the OJP, with the upper part of the plume head containingmost of the entrained eclogite pods, following the Yasuda et al. (1997) During eruption of Kwaimbaita-type lavas (Stage 1), a greater proportionof the entrained oceanic ex-crustal fragments is required to melt completely and mix with partial melts of the peridotite plume matrix. At thetime of eruption of Singgalo-type lavas (Stage 2) the few eclogitic fragments remaining melt completely and mix with partial melts of theperidotite matrix. (See text.)

31


higher amounts of fractionation, on average, as inferred CONCLUSIONSfor the Singgalo Fm. (relative to the Kwaimbaita). Over-

(1) Geochemical and geochronological study of 0·5 andall, although the crust of the OJP is much thicker3·5 km sections of pillowed and massive basaltic flows(30–43 km; e.g. Miura et al., 1996; Gladczenko et al.,composing crustal basement in central Malaita reveals a1997; Richardson et al., 2000) than that of the Caribbeansequence of 123·1 ± 1·4 Ma OJP crust resembling aPlateau (<20 km; Mauffret & Leroy, 1999), the grossgreatly expanded section of that at ODP Site 807. Thecrustal structure of the Caribbean and the Ontong Javaresults further confirm the importance of a majorPlateaux may be rather similar (e.g. Hussong et al., 1979;>122 Ma eruptive episode and a nearly plateau-wideKerr et al., 1997; Phinney et al., 1999).distribution of two isotopically distinct types of basalt inthe world’s largest oceanic plateau.Compositional structure of mantle source region

(2) As at Site 807, the two types of basalt form strati-The time difference between emplacement of the Kwaim- graphically separate groups. With the Kwaio Anticlinebaita and the Singgalo basalts is beyond the resolution as the type locality, we term the stratigraphically lowerof current 40Ar–39Ar dating capabilities, given the low group (maximum thickness >2·7 km) the KwaimbaitaK2O contents of most OJP basement rocks [but may just Fm. and the overlying group (maximum exposed thick-be distinguishable at Site 807 when biostratigraphic data ness of 750 m in central Malaita) the Singgalo Fm. Onare taken into account; see Larson & Erba (1999)]. the basis of closely similar isotopic and elemental char-Stratigraphically, the change in isotopic compositions is

acteristics, this nomenclature is extended to other loc-abrupt, and implies a source with two distinct com-ations where>122 Ma OJP basement has been sampled,positional regions and probably a component of verticalincluding northeastern Santa Isabel, Ramos Island, south-stratification.ern Malaita, ODP Site 807, and DSDP Site 289. TheIn Griffiths & Campbell’s (1990) model, hot plume-Singgalo magma type appears to be much more vo-source-derived material (light gray shading in Fig. 14a)luminous in the south than in the northern plateau (e.g.is carried to the stagnation point at the cap of the plumeonly 46 m of flows at Site 807 and none at Site 289).and spreads laterally into thin laminae. In a broad area

(3) Isotopic variation within each group is very limitedaround the center of the head, plume-source-dominatedand indicates an isotopically ocean-island-type mantleperidotite lies above a region enriched in entrainedsource with two distinct compositional regions, each ofperidotite (white in Fig. 14a). In this context, the Kwaim-which was fairly homogeneous relative to the scale ofbaita and Singgalo basalts can be interpreted as cor-melting. No evidence of normal-MORB-type mantle isresponding, respectively, to the upper, plume-source-present. The sharp stratigraphic boundary between thedominated zone and the underlying zone containingKwaimbaita and Singgalo formations in Malaita and,more entrained material. Griffiths & Campbell (1990)>1600 km away, between the two corresponding groupsalso argued that for a very large plume head, of whichat Site 807, coupled with the apparent lack of translationalthe OJP case is one of the two most viable examples,motion of this part of the Pacific plate between 125 andlittle entrainment of non-plume mantle would occur100 Ma, suggests a vertically stratified structure for thebecause the plume head would ascend too rapidly. ForOJP source mantle.this reason, the amount of plume-source material depicted

(4) The ocean-island-like isotopic signatures of OJPin Fig. 14a is much greater than that of entrained mantle.lavas can be accommodated with either a peridotiteFurthermore, Van Keken’s (1998) numerical simulationplume head (whether originating above a thermal bound-indicated that much of the plume-source material wouldary layer or at mid-mantle depths) or a composite diapirbe contained in the plume head and that the trailingmodel, in which a dominantly peridotitic plume head isconduit at this stage would be composed largely ofassumed to contain pockets of eclogite. Both models mayentrained non-plume mantle. Thus, Fig. 14a shows aalso accommodate the major and trace element data,reduced amount of plume-source material at greateralthough the compositions of model melts of a mixeddepth and tapping of mostly entrained mantle in theeclogite–peridotite source provide only rough fits to thelater stages of >122 Ma volcanism. In Yasuda et al.’sOJP data. Better fits are obtained assuming a two-(1997) composite diapir model, the upper part of thecomponent peridotitic source (Mahoney et al., 1993;plume head contains most of the eclogite pods, and Fig.Tejada et al., 1996; Neal et al., 1997; Tejada, 1998).14b shows most of them melting completely and mixingThe composite diapir model also requires an enormouswith partial melts of the peridotite matrix during theamount of eclogite to have been concentrated in aeruption of the Kwaimbaita basalts (Stage 1). At therelatively small part of the mantle if OJP magmas aretime of the Singgalo eruptions (Stage 2) fewer entrainedassumed to comprise a significant proportion of eclogiteeclogitic fragments remain, resulting in a decreased con-melt. Furthermore, the abrupt transition from Kwaim-tribution of melt derived from subducted oceanic crust

to the total melt. baita- to Singgalo-type basalts, and the reappearance

32


Bienvenu, P., Bougault, H., Joron, J. L., Treuil, M. & Dmitriev,some 30 my later of Kwaimbaita-like lavas are not readilyL. (1990). MORB alteration: rare-earth element/non-rare-earthexplained by the composite diapir model.hygromagmaphile element fractionation. Chemical Geology 82, 1–14.(5) In several important respects, the OJP fails to fit

Birkhold-VanDyke, A. L., Neal, C. R., Jain, J. C., Mahoney, J. J. &the predictions of, and thus poses a challenge to, any Duncan, R. A. (1996). Multi-stage growth for the Ontong Javaplume-head model: the absence of evidence for significant Plateau (OJP)? A progress report from San Cristobal (Makira),uplift despite construction of a crust of>35 km thickness Solomon Islands. EOS Transactions, American Geophysical Union 77,

F714.during what should have been the impact of the largestCastillo, P., Pringle, M. S. & Carlson, R. W. (1994). East Marianaplume head in the last 200 my, the lack of an attached

Basin tholeiites: Jurassic ocean crust or Cretaceous rift basalts relatedseamount chain (plume-tail trace), and the presence ofto the Ontong Java plume? Earth and Planetary Science Letters 123,at least two important episodes of eruptive activity sep- 139–154.

arated in time by >30 my. Chauvel, C., Hoffman, A. W. & Vidal, P. (1992). HIMU-EM: theFrench Polynesian connection. Earth and Planetary Science Letters 110,99–119.

Chen, C.-Y., Frey, F. A., Rhodes, J. M. & Easton, R. M. (1996).Temporal geochemical evolution of Kilauea volcano: comparisonACKNOWLEDGEMENTSof Hilina and Puna basalt. In: Basu, A. R. & Hart, S. (eds) EarthWe thank T. Babbs and A. Saunders for allowing us toProcesses: Reading the Isotopic Code. Geophysical Monograph, American Geo-

use their unpublished data, R. Magu for providing his physical Union 95, 161–181.field and petrographic notes, L. Kroenke, M. Garcia, J. Coffin, M. F. & Eldholm, O. (1994). Large igneous provinces: crustal

structure, dimensions, and external consequences. Reviews of GeophysicsSinton, A. Pietruszka, and G. Fitton for helpful dis-32, 1–36.cussions, K. Spencer, K. Rubin, T. Hulsebosch, and J.

Coffin, M. F. & Gahagan, L. M. (1995). Ontong Java and KerguelenC. Jain for help with the analytical work, and N. HulbirtPlateaux: Cretaceous Icelands? Journal of the Geological Society, Londonfor improving several figures. Ian Parkinson and an 152, 1047–1052.

anonymous reviewer provided valuable critical reviews. Coleman, P. J. & Kroenke, L. W. (1981). Subduction without volcanismThe Solomon Islands Ministry of Energy, Water, and in the Solomon Islands Arc. Geo-Marine Letters 1, 129–134.

Coleman, P. J., McGowran, B. & Ramsay, W. R. H. (1978). New,Mineral Resources (Geology Mapping) provided in-Early Tertiary ages for basal pelagites, Northeastern Santa Isabel,valuable logistical support, enthusiastic guidance, andSolomon Islands (central southwest flank, Ontong Java Plateau).diplomacy in the bush. This study was supported byBulletin of the Australian Society of Exploration Geophysicists 9(3), 110–114.NSF-EAR grants to J.J.M., C.R.N., and R.A.D., and a Cordery, M. J., Davies, G. F. & Campbell, I. H. (1997). Genesis of

Faculty Development Grant from the University of the flood basalts from eclogite-bearing mantle plumes. Journal of Geo-Philippines to M.L.G.T. physical Research 102, 20179–20197.

Cox, K. G. (1980). A model for flood basalt vulcanism. Journal of

Petrology 21, 629–650.Cox, K. G. (1993). Continental magmatic underplating. Philosophical

Transactions of the Royal Society of London, Series A 342, 155–166.REFERENCESCserepes, L. & Yuen, D. A. (2000). On the possibility of a second kind

Anderson, D. L. (1996). Enriched asthenosphere and depleted plumes. of mantle plume. Earth and Planetary Science Letters 183, 61–71.International Geology Reviews 38, 1–21. Davis, G. L. (1978). Zircons from the mantle. US Geological Survey Open

Anderson, D. L., Zhang, Y.-S. & Tanimoto, T. (1992). Plume heads, File Report 78–701, 86–88.continental lithosphere, flood basalts, and tomography. In: Storey, Duncan, R. A., Hooper, P. R., Rehacek, J., Marsh, J. S. & Duncan,B., Alabaster, T. & Pankhurst, R. J. (eds) Magmatism and Causes of A. R. (1997). The timing and duration of the Karoo igneous event,Continental Break-up. Geological Society, London, Special Publications 68, southern Gondwana. Journal of Geophysical Research 102, 18127–18138.99–124. Farnetani, C. G., Richards, M. A. & Ghiorso, M. S. (1996). Petrological

Arth, J. G. (1976). Behavior of trace elements during magmatic pro- models of magma evolution and deep crustal structure beneathcesses—a summary of theoretical models and their applications. hotspots and flood basalt provinces. Earth and Planetary Science LettersJournal of Research and Geological Survey 4, 41–47. 143, 81–96.

Babbs, T. L. (1997). Geochemical and petrological investigations of Galer, S. J. G. & Goldstein, S. L. (1996). Influence of accretion onthe deeper portions of the Ontong Java Plateau: Malaita, Solomon lead in the earth. In: Basu, A. R. & Hart, S. (eds) Earth Processes:Islands. Ph.D. dissertation, Leicester University, 254 pp. Reading the Isotopic Code. Geophysical Monograph, American Geophysical

Barron, A. J. M. (1993). The Geology of Northernmost Malaita: a Description Union 95, 75–98.of Sheets 8/160/7 and 8/150/8. Geological Division Report. Honiara: Garcia, M. O., Foss, D. J., West, H. B. & Mahoney, J. J. (1995).Solomon Islands Ministry of Natural Resources, 42 pp. Geochemical and isotopic evolution of Loihi volcano, Hawaii. Journal

Bedard, J. H. (1994). A procedure for calculating the equilibrium of Petrology 36(6), 1647–1674.distribution of trace elements among the minerals of cumulate rocks, Garcia, M. O., Rhodes, J. M., Trusdell, F. A. & Pietruszka, A. J.and the concentration of trace elements among the minerals of (1996). Petrology of lavas from Puu Oo eruption of Kilauea volcano:cumulate rocks, and the concentration of trace elements in the III. The Kupaianaha episode (1986–1992). Bulletin of Volcanology 58,coexisting liquids. Chemical Geology 118, 143–153. 359–379.

Berger, W. H., Kroenke, L. W., Mayer, L. A., et al. (eds) (1993). Ghiorso, M. S. & Sack, R. O. (1995). Chemical mass transfer inProceedings of the Ocean Drilling Program, Scientific Results, 130. College magmatic processes, IV. A revised and internally consistent thermo-

dynamic model for the interpolation and extrapolation of liquid–solidStation, TX: Ocean Drilling Program.

33


equilibria in magmatic systems at elevated temperatures and pressure. Jenner, G. A., Longerich, H. P., Jackson, S. E. & Fryer, B. J. (1990).ICP-MS—a powerful tool for high precision trace element analysis inContributions to Mineralogy and Petrology 119, 197–212.

Gladczenko, T. P., Coffin, M. F. & Eldholm, O. (1997). Crustal earth sciences: evidence from analysis of selected U.S.G.S. referencesamples. Chemical Geology 83, 133–148.structure of the Ontong Java Plateau: modeling of new gravity and

existing seismic data. Journal of Geophysical Research 102, 22711–22729. Kennedy, A. K., Lofgren, G. E. & Wasserburg, G. J. (1993). Anexperimental study of trace element partitioning between olivine,Govindaraju, K. (1994). 1994 Compilation of working values and

sample descriptions for 383 geostandards. Geostandards Newsletter, orthopyroxene, and melt in chondrules: equilibrium values andkinetic effects. Earth and Planetary Science Letters 115, 177–195.Special Issue 18, 15–16.

Green, T. H. (1994). Experimental studies of trace-element partitioning Kerr, A. C., Tarney, J., Marriner, G. F., Nivia, A. & Saunders, A. D.(1997). The Caribbean–Colombian Cretaceous igneous province:applicable to igneous petrogenesis—Sedona 16 years later. Chemical

Geology 117, 1–36. the internal anatomy of an oceanic plateau. In: Mahoney, J. J. &Coffin, M. F. (eds) Large Igneous Provinces: Continental, Oceanic, andGreen, T. H., Sie, S. H., Ryan, C. G. & Cousens, D. R. (1989). Proton

microprobe-determined partitioning of Nb, Ta, Zr, Sr, and Y Planetary Flood Volcanism. Geophysical Monograph, American Geophysical

Union 100, 123–144.between garnet, clinopyroxene and basaltic magma at high pressureand temperature. Chemical Geology 74, 210–216. Kogiso, T., Tatsumi, Y. & Nakano, S. (1997). Trace element transport

during dehydration processes in the subducted oceanic crust: 1.Griffiths, R. W. & Campbell, I. H. (1990). Stirring and structure inmantle starting plumes. Earth and Planetary Science Letters 99, 66–78. Experiments and implications for the origin of ocean island basalts.

Earth and Planetary Science Letters 148, 193–205.Grove, T. L., Kinzler, R. J. & Bryan, W. B. (1992). Fractionation ofmid-ocean ridge basalt (MORB). In: Phipps Morgan, J., Blackman, Kroenke, L. W. (1972). Geology of the Ontong Java Plateau. Hawaii Institute

of Geophysics, Report 72–75. Honolulu: University of Hawaii, 119 pp.D. K. & Sinton, J. (eds) Mantle Flow and Melt Generation at Mid-Ocean

Ridges. Geophysical Monograph, American Geophysical Union 71, 281–310. Kurz, M. D., Kenna, T., Kammer, D., Rhodes, J. M. & Garcia, M.O. (1995). Isotopic evolution of Mauna Loa volcano: a view fromGurenko, A. A. & Chaussidon, M. (1995). Enriched and depleted

primitive melts included in olivine from Icelandic tholeiites: origin the submarine southwest rift zone. In: Rhodes, J. M. & Lockwood,J. P. (eds) Mauna Loa Revealed: Structure, Composition, History, and Hazards.by continuous melting of a single mantle colum. Geochimica et Cosmo-

chimica Acta 58, 2905–2917. Geophysical Monograph, American Geophysical Union 92, 45–80.Lanphere, M. A., Dalrymple, G. B., Fleck, R. S. & Pringle, M. S.Hart, S. R. & Dunn, T. (1993). Experimental cpx/melt partitioning of

24 trace elements. Contributions to Mineralogy and Petrology 113, 1–8. (1990). Intercalibration of mineral standards for K–Ar and 40Ar–39Arage measurements. EOS Transactions, American Geophysical Union 71,Hauri, E. H., Wagner, T. & Grove, T. L. (1994). Experimental

partitioning of Th, U, Pb and other trace elements between garnet, 1658.Larson, R. L. & Erba, E. (1999). Onset of mid-Cretaceous greenhouseclinopyroxene and basaltic melts. Chemical Geology 117, 149–166.

Hawkins, M. P. & Barron, J. M. (1991). The Geology and Mineral Resources in the Barremian–Aptian: igneous events and the biological, sedi-mentary, and geochemical consequences. Paleoceanography 14, 663–of Santa Isabel. Geological Division Report J25. Honiara: Solomon Islands

Ministry of Natural Resources, 114 pp. 678.Libourel, G., Boivin, P. & Biggar, G. M. (1989). The univariant curveHerzberg, C. & O’Hara, M. J. (1998). Phase equilibrium constraints

on the origin of basalts, picrites, and komatiites. Earth-Science Reviews liquid = forsterite + anorthite + diopside in the system CMASat 1 bar: solid solutions and melt structure. Contributions to Mineralogy44, 39–79.

Herzberg, C. & Zhang, J. (1996). Melting experiments on anhydrous and Petrology 102, 406–421.Longhi, J. (1987). Liquidus equilibria and solid solution in the systemperidotite KLB-1: compositions of magmas in the upper mantle and

transition zone. Journal of Geophysical Research 101, 8271–8295. CaAl2Si2O8–Mg2SiO4–CaSiO3–SiO2 at low pressure. American Journal

of Science 287, 265–331.Hirose, H. & Kushiro, I. (1993). Partial melting of dry peridotites athigh pressures: determination of compositions of melts segregated Mahoney, J. J. (1987). An isotopic survey of Pacific oceanic plateaus:

implications for their nature and origin. In: Keating, B., Fryer, P.,from peridotite using aggregates of diamond. Earth and Planetary

Science Letters 114, 477–489. Batiza, R. & Boehlert, G. (eds) Seamounts, Islands, and Atolls. Geophysical

Monograph, American Geophysical Union 43, 207–220.Hughes, G. W. & Turner, C. C. (1976). Geology of Southern Malaita.Geological Survey Division, Bulletin 2. Honiara: Solomon Islands Mahoney, J. J. & Spencer, K. J. (1991). Isotopic evidence for the origin

of the Manihiki and Ontong Java oceanic plateaus. Earth and PlanetaryMinistry of Natural Resources, 80 pp.Hughes, G. W. & Turner, C. C. (1977). Upraised Pacific Ocean floor, Science Letters 104, 196–210.

Mahoney, J. J., Storey, M., Duncan, R. A., Spencer, K. J. & Pringle,southern Malaita, Solomon Islands. Geological Society of America Bulletin

88, 412–424. M. (1993). Geochemisty and geochronology of the Ontong JavaPlateau. In: Pringle, M., Sager, W., Sliter, W. & Stein, S. (eds) TheHussong, D. M., Wipperman, L. K. & Kroenke, L. W. (1979). The

crustal structure of the Ontong Java and Manihiki oceanic plateaus. Mesozoic Pacific. Geology, Tectonics, and Volcanism. Geophysical Monograph,

American Geophysical Union 77, 233–261.Journal of Geophysical Research 84, 6003–6010.Ito, G. T. & Clift, P. D. (1998). Subsidence and growth of Pacific Mahoney, J. J., Sinton, J. M., Kurz, M. D., MacDougall, J. D.,

Spencer, K. J. & Lugmair, G. W. (1994). Isotope and trace elementCretaceous plateaus. Earth and Planetary Science Letters 161, 85–100.Ito, G. T. & Taira, A. (2000). Compensation of the Ontong Java characteristics of a super-fast spreading ridge: East Pacific Rise,

13–23°S. Earth and Planetary Science Letters 121, 173–193.Plateau by surface and subsurface loading. Journal of Geophysical

Research 105, 11171–11183. Mahoney, J. J., Frei, R., Tejada, M. L. G., Mo, X. X., Leat, P. T. &Nagler, T. F. (1998). Tracing the Indian Ocean mantle domainJacobsen, S. B. & Wasserburg, G. J. (1980). Sm–Nd isotopic evolution

of chondrites. Earth and Planetary Science Letters 50, 139–155. through time: isotopic results from old West Indian, East Tethyan,and South Pacific seafloor. Journal of Petrology 39, 1285–1306.Jain, J. C. & Neal, C. R. (1996). Report of the inductively coupled

plasma-mass spectrometry (ICP-MS) facility, University of Notre Mahoney, J. J., Fitton, J. G., Wallace, P. J., et al. (eds) (2001). Proceedings

of the Ocean Drilling Program, Initial Reports, 192. College Station, TX:Dame 1993–1996. Open File Report, University of Notre Dame,30 pp. Ocean Drilling Program, in press.

34


Mauffret, A. & Leroy, S. (1999). Seismic stratigraphy and structure of and Solomon Islands arc. Journal of Geophysical Research 104, 20449–20466.the Caribbean igneous province. Tectonophysics 283, 61–104.

Rhodes, J. M. & Hart, S. R. (1995). Evolution of Mauna Loa volcano:Michael, P. J. (1999). Implications for magmatic processes at Ontongepisodic trace element and isotopic variations in historical MaunaJava Plateau from volatile and major element contents of CretaceousLoa lavas: implications for magma and plume dynamics. In: Rhodes,basalt glasses. Geochemistry, Geophysics, Geosystems 1, paper no.J. M. & Lockwood, J. P. (eds) Mauna Loa Revealed: Structure, Composition,1999GC000025 [10 170 words, 6 figures, 2 tables].History, and Hazards. Geophysical Monograph, American Geophysical UnionMiura, S., Shinohara, M., Takahashi, N., Araki, E., Taira, A., Suyehiro,92, 263–288.K., Coffin, M. F., Shipley, T. & Mann, P. (1996). OBS crustal

Richards, M. A., Duncan, R. A. & Courtillot, V. (1989). Flood basaltsstructure of the Ontong Java Plateau converging into Solomon islandand hot-spot tracks: plume heads and tails. Science 246, 103–107.arc (abstract). EOS Transactions, American Geophysical Union, 77F, 713.

Richardson, W. P., Okal, E. A. & Van der Lee, S. (2000). Rayleigh-McCulloch, M. T. & Black, L. P. (1984). Sm–Nd isotopic systematicswave tomography of the Ontong Java Plateau. Physics of the Earthof Enderby Land granulites and evidence for the redistribution ofand Planetary Interiors 118, 29–51.Sm and Nd during metamorphism. Earth and Planetary Science Letters

Rickwood, F. K. (1957). Geology of the island of Malaita. In: Geological71, 46–58.Reconnaissance of Parts of the Central Islands of B.S.I.P. Colonial GeologyMcDonough, W. F. & Sun, S.-s. (1995). The composition of the Earth.and Mineral Resources 6 (3), 300–306.

Chemical Geology 120, 223–253.Roden, M. F., Trull, T., Hart, S. R. & Frey, F. A. (1994). New He,McGinnis, C. E., Jain, J. C. & Neal, C. R. (1997). Characterization of

Nd, Pb, and Sr isotopic constraints on the constitution of thememory effects and development of an effective wash protocolHawaiian plume: results from Koolau Volcano, Oahu, Hawaii,

for the measurement of petrogenetically critical trace elements inUSA. Geochimica et Cosmochimica Acta 58(5), 1431–1440.

geological samples by ICP-MS. Geostandards Newsletter 21, 289–305.Sager, W. W. & Han, H.-C. (1993). Rapid formation of the Shatsky

McKenzie, D. & O’Nions, R. K. (1991). Partial melt distributions from Rise oceanic plateau inferred from its magnetic anomaly. Nature 364,inversion of rare earth element concentrations. Journal of Petrology 32, 610–613.1021–1091. Sheth, H. C. (1999). Flood basalts and large igneous provinces from

Moore, J. G., Normark, W. R. & Szabo, B. J. (1990). Reef growth and deep mantle plumes: fact, fiction, and fallacy. Tectonophysics 311,volcanism on the submarine southwest rift zone of Mauna Loa, 1–29.Hawaii. Bulletin of Volcanology 52, 375–380. Smith, A. D. & Lewis, C. (1999). The planet beyond the plume

Nakamura, Y. & Tatsumoto, M. (1988). Pb, Nd, and Sr isotopic hypothesis. Earth-Science Reviews 48, 135–182.evidence for a multicomponent source for rocks of Cook–Austral Smith, W. H. F. & Sandwell, D. T. (1996). New global seafloorIsland and heterogeneities of mantle plumes. Geochimica et Cosmochimica topography from satellite altimetry. EOS Transactions, American Geo-Acta 52, 2909–2924. physical Union 77, F315.

Neal, C. R., Mahoney, J. J., Kroenke, L. W., Duncan, R. A. & Stille, P., Unruh, D. M. & Tatsumoto, M. (1983). Pb, Sr, Nd, and HfPetterson, M. G. (1997). The Ontong Java Plateau. In: Mahoney, isotopic evidence of multiple sources for Oahu, Hawaii basalts. Nature

J. J. & Coffin, M. F. (eds) Large Igneous Provinces: Continental, Oceanic, 304, 25–29.Sun, S.-S. & McDonough, W. F. (1989). Chemical and isotopic sys-and Planetary Flood Volcanism. Geophysical Monograph, American Geophysical

tematics of oceanic basalts: implications for mantle composition andUnion 100, 183–216.processes. In: Saunders, A. D. & Norry, M. J. (eds) Magmatism in theNorrish, K. & Chappell, B. W. (1977). X-ray fluorescence spectrometry.Ocean Basins. Geological Society, London, Special Publications 42, 313–345.In: Zussman, J. (ed.) Physical Methods in Determinative Mineralogy. New

Tarduno, J. A., Sliter, W. V., Kroenke, L. W., Leckie, M., Mahoney,York: Academic Press, pp. 201–272.J. J., Musgrave, R. J., Storey, M. & Winterer, E. L. (1991). RapidPalacz, Z. & Saunders, A. D. (1986). Coupled trace element isotopeformation of the Ontong Java Plateau by Aptian mantle plumeenrichment on the Cook–Austral–Samoa Islands, southwest Pacific.volcanism Science 254, 399–403.Earth and Planetary Science Letters 79, 270–280.

Tatsumoto, M. (1978). Isotopic composition of lead in oceanic basaltParkinson, I. J., Arculus, R. J. & Duncan, R. A. (1996). Geochemistryand its implication to mantle evolution. Earth and Planetary Scienceof Ontong Java Plateau basalt and gabbro sequences, Santa Isabel,Letters 38, 63–87.Solomon Islands. EOS Transactions, American Geophysical Union 77,

Tejada, M. L. G. (1998). Geochemical studies of Pacific oceanicF715.plateaus: the Ontong Java Plateau and Shatsky Rise. Ph.D. dis-Peng, Z. X. & Mahoney, J. J. (1995). Drillhole lavas from the north-sertation, University of Hawaii, Honolulu, 328 pp.

western Deccan Traps, and the evolution of Reunion hotspot mantle.Tejada, M. L. G., Mahoney, J. J., Duncan, R. A. & Hawkins, M. P.

Earth and Planetary Science Letters 134, 169–185.(1996). Age and geochemistry of basement and alkalic rocks of

Petterson, M. G. (1995). The Geology of North and Central Malaita, Solomon Malaita and Santa Isabel, Solomon Islands, southern margin ofIslands (Including Implications of Geological Research on Makira, Savo Island, Ontong Java Plateau. Journal of Petrology 37(2), 361–394.Guadalcanal, and Choiseul between 1992 and 1995). Water and Mineral Todt, W., Cliff, R. A., Hanser, A. & Hoffman, A. W. (1996). EvaluationResources Division Geological Memoir 1/95. Honiara: Solomon of a 202Pb–205Pb double spike for high-precision lead isotope analysis.Islands Ministry of Energy, Water, and Mineral Resources. In: Basu, A. & Hart, S. (eds) Earth Processes: Reading the Isotopic Code.

Petterson, M. G., Neal, C. R., Mahoney, J. J., Kroenke, L. W., Geophysical Monograph, American Geophysical Union 95, 429–437.Saunders, A. D., Babbs, T. L., Duncan, R. A., Tolia, D., McGrail, Van Keken, P. (1998). Evolution of starting mantle plumes: a com-B. & Barron, M. (1997). Structure and deformation of north and parison between numerical and laboratory models. Earth and Planetarycentral Malaita, Solomon Islands: tectonic implications for the Science Letters 148, 1–11.Ontong Java Plateau–Solomon Arc collision, and for the fate of Vidal, P., Chauvel, C. & Brousse, R. (1984). Large mantle heterogeneityoceanic plateaus. Tectonophysics 283, 1–33. beneath French Polynesia. Nature 307, 536–538.

Phinney, E. J., Mann, P., Coffin, M. F. & Shipley, T. H. (1999). Walter, M. J. & Presnall, D. C. (1994). Melting behavior of simplifiedSequence stratigraphy, structure, and tectonic history of the south- lherzolite in the system CaO–MgO–Al2O3–SiO2–Na2O from 7 to

35 kbar. Journal of Petrology 35, 329–359.western Ontong Java Plateau adjacent to the North Solomon Trench

35


White, R. S. (1993). Melt production rates in mantle plumes. Philosophical systematics of Archean high-Mg lavas and modern oceanic basalts.Chemical Geology 126, 29–42.Transactions of the Royal Society of London 342, 137–153.

White, W. M. (1993). 238U/204Pb in MORB and open system evolution Yasuda, A., Fujii, T. & Kurita, K. (1997). A composite diapir modelfor extensive basaltic volcanism: magmas from subducted oceanicof the depleted mantle. Earth and Planetary Science Letters 115, 211–226.

Xie, Q ., McCuaig, T. C. & Kerrich, R. (1995). Secular trends in crust entrained within mantle plumes. Proceedings of the Japan Academy

73(10), 201–204.the melting depth of mantle plumes: evidence from HFSE/REE

36

basement geochemistry and geochronology of central malaita

Documents